These functions will get some additional callers in future revisions.
No functional change intended.
Sponsored by: Modirum MDPay
Sponsored by: Klara, Inc.
Details
Details
- Reviewers
glebius - Group Reviewers
network - Commits
- rG3e98dcb3d574: inpcb: Move inpcb matching logic into separate functions
Diff Detail
Diff Detail
- Repository
- rG FreeBSD src repository
- Lint
Lint Skipped - Unit
Tests Skipped - Build Status
Buildable 49753 Build 46643: arc lint + arc unit
Event Timeline
| sys/netinet/in_pcb.c | ||
|---|---|---|
| 2232 | We've been copying this XXX comment over and over. I think it originates from Robert Watson original SMP work. I don't think inp_vflag ever changes and the comment should be removed. | |
| sys/netinet6/in6_pcb.c | ||
| 998 | Maybe move the declaration to in_pcb_var.h to avoid declaring twice? | |
| sys/netinet/in_pcb.c | ||
|---|---|---|
| 2232 | It does change, see the IPV6_V6ONLY socket option handler, for example. In fact we had several legitimate bugs caused by unlocked checks of this flag. I think this particular instance of the warning can be removed: in the full patch set, we re-validate address fields and the vflag after the inpcb lock is acquired. | |
| sys/netinet6/in6_pcb.c | ||
| 998 | At one point these were slightly different: v4 has special logic for 4in6 sockets and I had an extra enum value for them. It might make sense to keep them separate for cases like that. | |