Details
- Reviewers
pauamma_gundo.com - Group Reviewers
network manpages - Commits
- rG02b958b19535: netlink: add netlink user documentation.
rG7366c0a49c9a: netlink: add netlink user documentation.
Diff Detail
- Repository
- rG FreeBSD src repository
- Lint
Lint Not Applicable - Unit
Tests Not Applicable
Event Timeline
Will review the rest later.
| share/man/man4/genetlink.4 | ||
|---|---|---|
| 32 | ||
| 43 | ||
| 44–45 | Maybe use "string family name" throughout instead? Having both a string identifier and another id(entifier) of an unspecified type is confusing. | |
| 46–47 | "the application" or "applications" (leaning toward the latter). | |
| 48 | ||
| 50 | ||
| 53 | Not sure "families" is right, but there's definitely a word missing. | |
| 64–66 | Or remove "The". | |
| 69 | Same as above. | |
| 71 | ||
| 73 | Needs to say which family id it has, here or somewhere. Here is probably best. | |
| 81 | What I think you mean. | |
Done for this round.
| share/man/man4/netlink.4 | ||
|---|---|---|
| 57 | VNETs? Not other kinds of network? | |
| 68 | ||
| 84 | ||
| 112 | Spurious space. | |
| 276 | ||
| share/man/man4/rtnetlink.4 | ||
| 104 | ||
| 105 | ||
| 110 | VNET? | |
| 135–136 | ||
| 222 | Or "The required objects are" if more than 1. | |
| 290 | ||
| 293 | Is this really restricted to VNETs? | |
| 296 | ||
| 309 | For brevity and consistency with later ones. | |
| 311 | ||
| 341 | Is there an active 802.1x standard other than 802.11? | |
| 421–423 | US spelling | |
| 437 | (or get rid of "The") | |
| 454 | ||
| 463 | ||
| 470 | ||
| 476 | Also, VNET? | |
| 492 | ||
| 494 | ||
| 504 | ||
| share/man/man4/netlink.4 | ||
|---|---|---|
| 236 | NETLINK_DROP_MEMBERSHIP | |
| share/man/man4/genetlink.4 | ||
|---|---|---|
| 73 | GENL_ID_CTRL is the actual Id. I’d rather not specify its value here, to avoid any encouragement for using number instead of a header constant. Also typically we don’t provide such mappings in man page. Thoughts? | |
| share/man/man4/netlink.4 | ||
| 57 | Vnets. Or namespaces in Linux land | |
| share/man/man4/rtnetlink.4 | ||
| 110 | Yes. | |
| 293 | Well, yes. Similar to ifconfig. There is a socket option & attribute that allows to interact with the objects in the VNET different from the current one, but it’s not implemented yet. | |
| 341 | A lot (even too many, maybe). 802.1D is about bridging, including loop prevention protocols, .1Q is about vlans and so on. 802.1X is a standard handling authentication in LAN / wireless. If network is using it, there will be a temporary state when an interface is already able to send/receive frames, but encryption is not negotiated yet. This condition is an example of the aforementioned state. | |
Other than a few nits all fixable on commit, LGTM. If it matches the implementation, it's good to go.
| share/man/man4/genetlink.4 | ||
|---|---|---|
| 28 | Bump on commit. | |
| 73 | Never mind. I misparsed the whole paragraph. Ignore the noise. | |
| share/man/man4/netlink.4 | ||
| 112 | "32-bit" when the length is in bytes doesn't look quite right. For consistency, I'd change the other occurrences as well. | |
| 200 | "message", maybe? | |
| 277 | ||
| 341 | ||
| share/man/man4/netlink.4 | ||
|---|---|---|
| 341 | Mm, I'm a bit unsure about this one. Technically it's indeed an OS feature, but it sounds a bit abstract to me. We typically say, "X has IPv6 [protocol] support" or "X has TCP [protocol]" and refer to the word "feature" when there is no better word to describe it. Thoughts? | |
| share/man/man4/netlink.4 | ||
|---|---|---|
| 341 | Well, there's definitely a word missing, so we need to choose one to add. I was trying to avoid repeating "protocol", but that would work. | |