svn commit: r205104 - in head/sys: dev/xen/netback netinet netinet6

[Randall Stewart]

Hmm.

did not think of that.. we COULD possible do it another way.. a bit  
harder
but possible.. i.e. have the delayed sack code actually look into
the mbufs and see if its ipv4 or ipv6.. I thought about doing it
that way but it takes more cycles ;-o

I could refactor that this way if you want... it would mean a few more  
de-ref's and
looking to see if its a v4 or v6 packet and then doing the proper  
offset...

not to bad but awkward ;-0

R

On Mar 13, 2010, at 9:58 PM, Robert Watson wrote:

>
> On Fri, 12 Mar 2010, Randall Stewart wrote:
>
>> The proper fix for the delayed SCTP checksum is to
>> have the delayed function take an argument as to the offset
>> to the SCTP header. This allows it to work for V4 and V6.
>> This of course means changing all callers of the function
>> to either pass the header len, if they have it, or create
>> it (ip_hl << 2 or sizeof(ip6_hdr)).
>
> It will become harder to MFC SCTP changes like this in the future,  
> BTW.  It's only good (bad) luck that this change doesn't break the  
> KBI depended on by the ipdivert module.  If SCTP had been enabled in  
> the module (which it probably should be, with options SCTP in  
> GENERIC), then the module would be broken, meaning that we couldn't  
> MFC this change as-is.
>
> After MFCing this change, I recommend fixing ipdivert, but also  
> increased caution in changing kernel interfaces that may be depended  
> on by modules...
>
> Robert N M Watson
> Computer Laboratory
> University of Cambridge
>

------------------------------
Randall Stewart
803-317-4952 (cell)
803-345-0391(direct)