Page MenuHomeFreeBSD

tcp: stop spurious rescue retransmissions
ClosedPublic

Authored by rscheff on Apr 25 2021, 11:03 AM.

Details

Summary

As stale data from a prior SACK loss recovery episode persists,
a subsequent reordering event may trigger a spurious rescue
retransmission, which in turn misbehaves due to other stale
data.

To prevent sending an erraneous rescue retransmission, clear
the sackhint.delivered_data value, once finished with a SACK
loss recovery.

Furthermore, initialize snd_fack to th_ack on a partial ack,
to pull it forward from whereever the last proper SACK may
have been.

Test Plan

Perform long-duration testing with a low background of packet
reordering and very few losses (or sufficient reordering for
dupthresh to be exceeded).

Reported by: pho@
MFC after: 3 days
Sponsored by: NetApp, Inc.

Diff Detail

Repository
rG FreeBSD src repository
Lint
Automatic diff as part of commit; lint not applicable.
Unit
Automatic diff as part of commit; unit tests not applicable.