>Number:         10617
>Category:       kern
>Synopsis:       sysctl(3) implies we support SACK
>Confidential:   no
>Severity:       non-critical
>Priority:       medium
>Responsible:    kern-bug-people
>State:          open
>Class:          doc-bug
>Submitter-Id:   net
>Arrival-Date:   Mon Jul 17 12:37:01 PDT 2000
>Closed-Date:
>Last-Modified:
>Originator:     John Hawkinson
>Release:        netbsd-1-5 of 16 July 2000
>Organization:
MIT
>Environment:
	
System: NetBSD zorkmid.mit.edu 1.5B NetBSD 1.5B (ZORKMID-$Revision: 1.16 $) #7: Fri Jul 14 22:51:02 EDT 2000 jhawk@zorkmid.mit.edu:/usr/local/netbsd-current/src/sys/arch/i386/compile/ZORKMID i386


>Description:
	sysctl(3) implies we support SACK:
             tcp.sack
                     Returns a value which determines the level of Selective
                     Acknowledgement supported by TCP.  If 2, we will transmit
                     and receive SACK options.  If 1, we will transmit SACK
                     options, but ignore any SACK options received.  If 0,
                     SACK is disabled.
	Apparently we don't; weren't we supposed to have SACK "RSN" in 1998?

>How-To-Repeat:
	Observe that tcp_do_sack defaults to 1 (rx) and not 2 (tx+rx),
	and wonder what's going on.
>Fix:
	Implement, implement, implement?

	At least reword te sysctl(3) text to clarify what we do and do
	not support. I'm not really clear on that, so I have not
	offered any text. Perhaps verbiage in tcp(4) would also be
	appropriate (perhaps not if it doesn't work)? 
>Release-Note:
>Audit-Trail:
>Unformatted: