[Bug 275798] panic: sackhint bytes rtx >= 0
- Go to: [ bottom of page ] [ top of archives ] [ this month ]
Date: Sun, 17 Dec 2023 15:29:38 UTC
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=275798
Richard Scheffenegger <rscheff@freebsd.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
Assignee|bugs@FreeBSD.org |rscheff@freebsd.org
--- Comment #4 from Richard Scheffenegger <rscheff@freebsd.org> ---
Can you please enable bblogging:
Enable BBR logging for all TCP sessions:
# sysctl net.inet.tcp.bb.log_auto_mode=4
# sysctl net.inet.tcp.bb.log_auto_all=1
# sysctl net.inet.tcp.bb.log_auto_ratio=1
when this panic is easily reproducible?
Alternatively, a tcpdump of the tcp session(s) within the ipsec tunnel (not a
interface packet capture, but after the decoding).
The core should then contain full chains of the relevant information we could
extract offline...
The full sequence of tcp packets from the onset of packet loss and the ACKs
(with SACKs) from the remote side should yield some clue why this is happening.
This appears to be some kind of SACK (selective acknowledgment) accounting
error - but no recent change in this area...
--
You are receiving this mail because:
You are on the CC list for the bug.