On Thu, Oct 13, 2005 at 11:15:51AM -0400, Andrew Atrens wrote:
>
> Not sure what these 'errno 209's are about either :('
i think i saw those in 3.7 and didn't have them affect anything
i was doing adversely, at least noticably so...
they aren't showing in 3.8/oct.2 ( or sep27 )
> # ipsecadm group -cpi 1004 -spi 1004 -proto ipcomp -spi2 1000 -proto2 esp
> -dst 47.128.22.193 -dst2 47.128.22.193
>
> ipsecadm: write: Invalid argument
i'm getting this also, i duplicated the info you have from the 'ipsecadm
show',
and then i get the 'write invalid' with all the permutations of 'group' i try.
i threw a bunch of 'fprintf' debug markers into ipsecadm.c during the
case GRP_SPI at the bottom, it makes it through them all, does the break
and dies during "xf_set(iov, cnt, smsg.sadb_msg_len * 8);"
tried doing a printf of those values before the xf_set, but using '%s' gave
me a segfault. "%X %X %X" gives "CFBC736C 8 50" for the same line you
have right above.
trying with ah+esp ( 47.128.22.193/esp/1000, 47.128.22.193/ah/1002) gave
"CFBC1250 8 50"; i notice the first one, the 'iov', that is changing every
time.
flailed around some more, looked at the case statement up top, it looked
almost like the comment for "SA header" should be "SA2 header" ?
for the second part ( eg, there is a "Destination2 address header" ), but
the actual variables in it seemed to talk about sa2 ( or "8", rather,
but that looks like the second one ). noticed that there is a 'protocol'
section that happens at the end, but only one of them; i copied the iov
stuff ( and the smsg one ) up before the 2nd SA header part, but that
didn't change; also tried commenting out the '=sproto2' line and made
the current protocol stuff say protocol2; it still compiled, and when
i ran the ipsecadm group, that junky output i got from before was now
like "CFBD332C 9 58"; so 8 -> 9 and 50 -> 58... :/
but perhaps all of that was entirely useless and i'm doing the group
thing wrong too
jared
--
[ openbsd 3.8 GENERIC ( sep 27 ) // i386 ]
