On Sun, Jun 29, 2003 at 01:02:58AM +1000, Darren Reed wrote: > Couple of pointers...
Always appreciated!
> The routines that implement the "put" ioctl expect to be copying data
> from userspace, so you'll need to account for that.
Yesh -- this is a userland proxy. =)
The reason I was asking is because the current routines don't do a
lot of checking of incoming entries, so I was playing Whack-A-Mole(tm)
finding code referring to certain members without first making sure
they weren't NULL. (E.g., fr_nat{in,out} deref np/nat->nat_ptr without
first making sure they're not NULL, even though fr_natputent *does*
test this before inserting the entry and returning w/o error.)
I wasn't sure if these routines were for general, public consumption
or were restricted to entries previously copied from the kernel (as
a backup before shutdown?)
> Otherwise, I can't see why not.
Neat. Thanks again.
--
ryan beasley <[EMAIL PROTECTED]>
GPG ID: 0x16EFBD48
pgp00000.pgp
Description: PGP signature
