Page MenuHomeFreeBSD
Differential D35387

Sanity check of SACK holes on retransmissions
ClosedPublic
Actions

Authored by rscheff on Jun 2 2022, 4:55 PM.
Tags
None
Referenced Files
F166652226: D35387.id106711.diff
Sat, Aug 15, 5:48 AM
F166567152: D35387.id106580.diff
Fri, Aug 14, 1:41 PM
Unknown Object (File)
Wed, Aug 12, 5:18 PM
Unknown Object (File)
Wed, Aug 12, 5:17 PM
Unknown Object (File)
Wed, Aug 12, 5:17 PM
Unknown Object (File)
Wed, Aug 12, 5:17 PM
Unknown Object (File)
Wed, Aug 12, 4:54 PM
Unknown Object (File)
Wed, Aug 12, 11:42 AM
View All Files
Subscribers
glebius
imp
melifaro

Details

Reviewers
• hselasky
tuexen
glebius
Group Reviewers
transport
Commits
rG91d6afe6e2a9: tcp: Sanity check of SACK holes on retransmissions
Summary

Adding a few KASSERT() to validate sanity of sack holes, and bail out if sack hole is inconsistent to avoid panicing non-invariant builds.

Diff Detail

Repository
rS FreeBSD src repository - subversion
Lint
Lint Passed
Unit
No Test Coverage
Build Status
Buildable 45821
Build 42709: arc lint + arc unit

Event Timeline

rscheff created this revision.Jun 2 2022, 4:55 PM
Herald added a reviewer: transport. · View Herald TranscriptJun 2 2022, 4:55 PM
Herald added subscribers: glebius, melifaro, imp. · View Herald Transcript
rscheff requested review of this revision.Jun 2 2022, 4:55 PM
Harbormaster completed remote builds in B45819: Diff 106575.Jun 2 2022, 4:55 PM
• hselasky added inline comments.Jun 2 2022, 5:24 PM
sys/netinet/tcp_sack.c
968

Don't you need to use the SEQ macros for this?

glebius added inline comments.Jun 2 2022, 6:45 PM
sys/netinet/tcp_sack.c
976

Maybe print something in non-debug case?

rscheff updated this revision to Diff 106580.Jun 2 2022, 6:49 PM
  • use SEQ_LT / GEQ macros
Harbormaster completed remote builds in B45821: Diff 106580.Jun 2 2022, 6:49 PM
• hselasky accepted this revision.Jun 2 2022, 6:51 PM
rscheff updated this revision to Diff 106581.Jun 2 2022, 6:59 PM
  • add logging
Harbormaster completed remote builds in B45822: Diff 106581.Jun 2 2022, 6:59 PM
glebius accepted this revision.Jun 2 2022, 7:01 PM
tuexen added inline comments.Jun 2 2022, 8:32 PM
sys/netinet/tcp_sack.c
976

When printing them using %d, don't some of them show up as negative numbers? Would %u make more sense?

rscheff updated this revision to Diff 106711.Jun 7 2022, 6:49 AM
  • use SEQ_LT / GEQ macros
  • add logging
Harbormaster completed remote builds in B45891: Diff 106711.Jun 7 2022, 6:49 AM
rscheff updated this revision to Diff 106712.Jun 7 2022, 7:04 AM
  • change format to %u
Harbormaster completed remote builds in B45892: Diff 106712.Jun 7 2022, 7:04 AM
rscheff marked an inline comment as done.Jun 7 2022, 7:04 AM

Thanks. Commiting with 1 week MFC...

rscheff marked 2 inline comments as done.Jun 7 2022, 7:04 AM
This revision was not accepted when it landed; it landed in state Needs Review.Jun 7 2022, 7:41 AM
Closed by commit rG91d6afe6e2a9: tcp: Sanity check of SACK holes on retransmissions (authored by rscheff). · Explain Why
This revision was automatically updated to reflect the committed changes.
rscheff added a commit: rG91d6afe6e2a9: tcp: Sanity check of SACK holes on retransmissions.
rscheff mentioned this in D35423: tcp: skip sackhole checks on NULL.Jun 7 2022, 6:20 PM
rscheff mentioned this in rG231e0dd5d1fb: tcp: skip sackhole checks on NULL.Jun 7 2022, 6:47 PM

Revision Contents
Changeset List

PathSize
sys/
netinet/
9 lines

Diff 106580

View Options

sys/netinet/tcp_sack.c

Loading...