The "cgdconfig" program was much harder to build. Wanted the tools directory, 
which wants, the gnu directory, which wants... but maybe I had it configed to do 
cross-compile. Anyway, that built as well.

One thing to add to the patch file is perhaps "/dev/MAKEDEV".

All working great, great work. Out of curiousity, why is the patch full of 
RCS/*,v files, are they useful generally, in this situation? (I know what they 
are, just wondered what use they have in a patch - probably something obvious).

Lund


Jorgen Lundman wrote:
> 
> I have a need for this on a new system, and didn't feel like running 
> current, so I grabbed the 1.6.1 sources, and applied the cgd patch.
> 
> I found the following issues after patching:
> 
> ./../../../crypto/des/des.h:68: parse error before `u_int32_t'
> 
> To fix it so it compiled (probably not the correct solution..) I added:
> 
> #include <sys/types.h> to that file.
> 
> 
> Then I always have issues with bf_cbc.c, even when I did my own patched 
> is 1.5.3. It just does not compile out of plain kernel sources. Are we 
> even supposed to have a openssl/blowfish.h for kernel? Anyway, I change 
> it to:
> 
> #include <sys/types.h>
> #include "blowfish.h"  instead of <openssl/blowfish.h>
> 
> Also, all 4 calls
>                         BF_encrypt(tin,(BF_KEY *)schedule);
> 
> since it removes "const" and -Werror interfers.
> 
> 
> However, I then get to:
> 
> dev/cgd.c:90: `DTYPE_CGD' undeclared here (not in a function)
> 
> Which is not mentioned in the patch apart from that .c file. I can take 
> a guess that using "16" is free in my disklabel.h but would be nice to 
> be sure. That makes it all compile at least. Not tried booting it yet.
> 
> 
> Lundy
> 
> 

-- 
Jorgen Lundman       | <lundman@lundman.net>
Unix Administrator   | +81 (0)3 -5456-2687 ext 1017 (work)
Shibuya-ku, Tokyo    | +81 (0)90-5578-8500          (cell)
Japan                | +81 (0)3 -3375-1767          (home)