git: 91ee2d8d9a67 - main - tcp: clear SACK state when scoreboard is forcefully freed

The branch main has been updated by rscheff:

URL: https://cgit.FreeBSD.org/src/commit/?id=91ee2d8d9a67c4b70612cb02d6d8bcb6a226cb6f

commit 91ee2d8d9a67c4b70612cb02d6d8bcb6a226cb6f
Author:     Richard Scheffenegger <rscheff@FreeBSD.org>
AuthorDate: 2023-10-11 20:38:34 +0000
Commit:     Richard Scheffenegger <rscheff@FreeBSD.org>
CommitDate: 2023-10-11 21:09:58 +0000

    tcp: clear SACK state when scoreboard is forcefully freed
    
    When a Retransmission Timeout happens during an on-going SACK loss recovery
    episode, the internal SACK accounting was not cleared.
    
    Reported by: pho
    Reviewed By: tuexen, #transport
    Sponsored by: NetApp, Inc.
    Differential Revision: https://reviews.freebsd.org/D42162
---
 sys/netinet/tcp_sack.c | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/sys/netinet/tcp_sack.c b/sys/netinet/tcp_sack.c
index 9a01ff684cbb..8647630bb6bc 100644
--- a/sys/netinet/tcp_sack.c
+++ b/sys/netinet/tcp_sack.c
@@ -888,6 +888,10 @@ tcp_free_sackholes(struct tcpcb *tp)
 	while ((q = TAILQ_FIRST(&tp->snd_holes)) != NULL)
 		tcp_sackhole_remove(tp, q);
 	tp->sackhint.sack_bytes_rexmit = 0;
+	tp->sackhint.sacked_bytes = 0;
+	tp->sackhint.delivered_data = 0;
+	tp->sackhint.lost_bytes = 0;
+	tp->sackhint.hole_bytes = 0;
 
 	KASSERT(tp->snd_numholes == 0, ("tp->snd_numholes == 0"));
 	KASSERT(tp->sackhint.nexthole == NULL,