Differential D49628

rtwn: don't treat UDP/TCP checksum failure as permanent failure
ClosedPublic
Actions

Authored by adrian on Apr 2 2025, 12:00 AM.

Details

Reviewers

jrtc27

Group Reviewers

wireless

Commits

rGd5f5193e2db9: rtwn: don't treat UDP/TCP checksum failure as permanent failure

Summary

jrtc27@freebsd.org reported that DHCP wasn't working on some
networks. She dug into it and found that the RTL8812AU/RTL8812AU
NICs seem to be failing UDP frames w/ a zero checksum, which is
a valid "there's no checksum" checksum.

So, just pass those frames up the stack and let the IP stack
deal with it. If the hardware claims the frames did pass TCP/UDP
checksum then still mark those frames with the checksum offload
bits.

PR: kern/285387

Diff Detail

Repository

rG FreeBSD src repository

Lint

Lint Not Applicable

Unit

Tests Not Applicable

Event Timeline

adrian created this revision.Apr 2 2025, 12:00 AM

Herald added a subscriber: imp. · View Herald TranscriptApr 2 2025, 12:00 AM

adrian requested review of this revision.Apr 2 2025, 12:00 AM

Harbormaster completed remote builds in B63273: Diff 153007.Apr 2 2025, 12:00 AM

adrian added a parent revision: D49596: rtwn: enable reception of BAR frames.Apr 2 2025, 12:01 AM

adrian added reviewers: wireless, jrtc27.

adrian added a project: wireless.

olce mentioned this in rGf01d26dec67f: MAC/do: Rules: <from> and <to> parts now to be separated by '>'.Apr 2 2025, 4:27 PM

olce mentioned this in rG8047b85cbe17: MAC/do: Rules: <from> and <to> parts now to be separated by '>'.Apr 8 2025, 1:42 PM

adrian added a child revision: D49764: net80211: refactor sequence number assignment code.Apr 10 2025, 8:36 PM

bz added a subscriber: bz.Apr 12 2025, 1:00 AM

bz added inline comments.

sys/dev/rtwn/rtl8812a/r12a_rx.c
202	If this is just UDP you could filter those errors out if I read the above debug printf correctly. Slight optimization...?

adrian added inline comments.Apr 12 2025, 1:52 PM

sys/dev/rtwn/rtl8812a/r12a_rx.c
202	I thought about only returning 0 for UDP + CKSUM error, but -1 for TCP + CKSUM error. @jrtc27 wanna try changing the logic locally and make sure you don't see any weird corner cases?