Page MenuHomeFreeBSD
Differential D49818

net80211: move the storing of ni in the mbuf further down
Changes PlannedPublic
Actions

Authored by bz on Apr 13 2025, 11:13 PM.
Tags
None
Referenced Files
Unknown Object (File)
Mon, Dec 15, 11:00 AM
Unknown Object (File)
Wed, Nov 26, 1:26 PM
Unknown Object (File)
Tue, Nov 25, 3:59 PM
Unknown Object (File)
Nov 12 2025, 8:04 AM
Unknown Object (File)
Nov 10 2025, 12:58 AM
Unknown Object (File)
Nov 8 2025, 8:10 AM
Unknown Object (File)
Nov 7 2025, 9:43 PM
Unknown Object (File)
Nov 6 2025, 1:57 AM
View All 46 Files
Subscribers
adrian
glebius
imp
melifaro

Details

Reviewers
None
Group Reviewers
wireless
Summary

We store the node (ni) in the mbuf meta data already after
ieee80211_dwds_mcast() (see comment). However ieee80211_encap()
further down can possibly allocate a new mbuf calling into
ieee80211_mbuf_adjust() so move storing the node information right
in front of the call which will rely on it to avoid any further
problems.

Sponsored by: The FreeBSD Foundation
MFC after: 3 days

Diff Detail

Repository
rG FreeBSD src repository
Lint
Lint Passed
Unit
No Test Coverage
Build Status
Buildable 63491
Build 60375: arc lint + arc unit

Event Timeline

bz created this revision.Apr 13 2025, 11:13 PM
Herald added subscribers: glebius, melifaro, imp. · View Herald TranscriptApr 13 2025, 11:13 PM
bz requested review of this revision.Apr 13 2025, 11:13 PM
Harbormaster completed remote builds in B63491: Diff 153623.Apr 13 2025, 11:13 PM
adrian added a subscriber: adrian.Apr 13 2025, 11:28 PM

Is that true though? like, the superg/amsdu transmit path is also queuing the packet and it needs the node reference for it to do so.

bz planned changes to this revision.Apr 13 2025, 11:37 PM
In D49818#1135726, @adrian wrote:

Is that true though? like, the superg/amsdu transmit path is also queuing the packet and it needs the node reference for it to do so.

Sugar. And ff_transmit() has the same problem that it might lose it and then we go kaboom in ieee80211_parent_xmitpkt error case; okay; there's also an improvement to be had there. I'll do the other one first and revise this. I think the answer is to audit ieee80211_encap callers and if all do the same make sure ieee80211_encap preserves the meta information.

adrian added a comment.Apr 14 2025, 12:06 AM

Yeah, let's both do that (some auditing/review) first.

My earlier diff on documenting the transmit paths at least document the node ref stuff, I can also go document the encap behaviour too in another diff.

Revision Contents
Changeset List

PathSize
sys/
net80211/
16 lines

Diff 153623

View Options

sys/net80211/ieee80211_output.c

Loading...