Re: kern/50417: TCP_CHECKSUM panic()

[christos%zoulas.com@localhost (Christos Zoulas)]

On Nov 10,  4:05pm, gergely%egervary.hu@localhost (gergely%egervary.hu@localhost) wrote:
-- Subject: kern/50417: TCP_CHECKSUM panic()

| >Number:         50417
| >Category:       kern
| >Synopsis:       TCP_CHECKSUM panic()
| >Confidential:   no
| >Severity:       serious
| >Priority:       low
| >Responsible:    kern-bug-people
| >State:          open
| >Class:          sw-bug
| >Submitter-Id:   net
| >Arrival-Date:   Tue Nov 10 16:05:00 +0000 2015
| >Originator:     Gergely EGERVARY
| >Release:        NetBSD 6.1_STABLE
| >Organization:
| >Environment:
| NetBSD galileo.poli.hu 6.1_STABLE NetBSD 6.1_STABLE (GALILEO) #7: Tue Nov 10 14:00:54 CET 2015  root%venus.poli.hu@localhost:/usr/src/sys/arch/amd64/compile/GALILEO amd64
| >Description:
| Compiling TCP_CHECKSUM into kernel, and starting a checksum-enabled TCP session (bgpd from quagga) panics kernel. It's 100% reproducible for me.
| 
| Is TCP_CHECKSUM known to be broken in 6.x?

You mean TCP_SIGNATURE? This was fixed in:

revision 1.178
date: 2014-10-25 11:07:13 -0400;  author: christos;  state: Exp;  lines: +30 -25;  commitid: ihyehwegSyD1TAVx;
Avoid stack overflow when SACK and TCP_SIGNATURE are both present. Thanks
to Jonathan Looney for pointing this out.

christos