unix/dgram: add a specific receive method - uipc_soreceive_dgram
ClosedPublic
Actions

Authored by glebius on May 23 2022, 9:01 PM.

Details

Reviewers

Group Reviewers

network
transport

Commits

rGe3fbbf965e94: unix/dgram: add a specific receive method - uipc_soreceive_dgram

Summary

With this second step PF_UNIX/SOCK_DGRAM has protocol specific
implementation. This gives some possibility performance
optimizations. However, it still operates on the same struct
socket as all other sockets do.

Diff Detail

Repository

rG FreeBSD src repository

Lint

Lint Not Applicable

Unit

Tests Not Applicable

Event Timeline

glebius created this revision.May 23 2022, 9:01 PM

Herald added a subscriber: imp. · View Herald TranscriptMay 23 2022, 9:01 PM

glebius requested review of this revision.May 23 2022, 9:01 PM

Harbormaster completed remote builds in B45677: Diff 106295.May 23 2022, 9:01 PM

glebius added a parent revision: D35295: unix/dgram: cleanup uipc_send of PF_UNIX/SOCK_DGRAM, step 2.May 23 2022, 9:01 PM

glebius added a child revision: D35297: unix/dgram: inline sbappendaddr_locked() into uipc_sosend_dgram().

glebius added a child revision: D35305: libc/syslog: deprecate use of "/var/run/logpriv".May 23 2022, 9:07 PM

glebius added reviewers: network, markj.May 23 2022, 9:11 PM

Note: this patch is a step towards D35303.

markj added inline comments.May 24 2022, 2:50 PM

sys/kern/uipc_usrreq.c
1290	Doesn't this copy internalized control messages?
1336
1445	Looks like this error is lost?

glebius marked 3 inline comments as done.May 25 2022, 1:33 AM

glebius added inline comments.

sys/kern/uipc_usrreq.c
1290	Yes, and this is the intent. Before we did the same in soreceive_generic() uipc_socket.c:2159
1445	Good catch! In the soreceive_generic() we catch the error much later and my code doesn't have it. I'm looking in making unp_externalize() to never fail, at least for SOCK_DGRAM.

Simplify and improve handling of unp_externalize().

Harbormaster completed remote builds in B45708: Diff 106359.May 25 2022, 7:35 AM

markj accepted this revision as: markj.May 25 2022, 7:49 PM

markj added inline comments.

sys/kern/uipc_usrreq.c
1434
1435	Related, unp_scan() will visit `m->m_nextpkt`, since it was intended to operate on the entire socket buffer. I can't find any problems caused by that, but it seems rather subtle. Might be worth asserting `m->m_nextpkt == NULL` before calling unp_scan() below and elsewhere in your patch set. Or improve the KPI to remove such footguns.