--=-=-=
Content-Transfer-Encoding: quoted-printable


Andrew Brown <atatat@atatdot.net> writes:

>>> is there any easy way of disabling it?  or making an lkm that will
>>> load more "easily"?
>>
>>How do you determine what's 'safe' and what not? Once you
>>dereference any kernel struct pointer (e.g. struct proc or struct lwp),
>>there is incompatibility. Obviously it's not possible for a LKM
>>to do anything useful and be ABI safe.
>
> i'll buy that, but i'm not *doing* any of that.  at all.  that's my
> point.  i can completely pigeonhole the api that i'm using.
[...]
> maybe i'll just make my own kernel option to disable the check...

So when struct sysent change, you will fall over...

I think you are complaning about the wrong thing. The problem I see is that
the version number is not related (well maybe, is it) to the kernel ABI for
release branches, while its most certainly is for current/HEAD.

Love


--=-=-=
Content-Type: application/pgp-signature

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.3 (NetBSD)

iQEVAwUAP8tLq3W+NPVfDpmCAQIojgf/e98G9gwidwhpvY2F0pYtETbWNGfLhk8R
lSHOnrXXNtdO7dk+jseTv+9isAUmdwl7S3vzvm+RdwqifeOPxMQOflP5xbkFcmho
PzBqwuMU2wJcsdMa8r+uOcFwOEfih9sEBqXE+8qz9XKjFum7pqq7smoQo2pl0H5/
iIRz/1Yd6FIDKXXuu4Cd5ga5iW55slWxlcjt0RUwaeyFZALj1+NWuadiXfqnaSqY
vrNtu4YNDStuLbyJQ6YaR1nvfjd9nzRDrRytRMsZ/P7rh0wngBsbcE/45NlWr7Cv
E31yN/bv4OoB4EoVvhbRyCD2EGUE2VuV+X4Av3wb4/bI6xN7xSYneg==
=X28x
-----END PGP SIGNATURE-----
--=-=-=--