In article <20020304172127.27532.qmail@mail.netbsd.org>
eeh@netbsd.org wrote:

> | > If not, what type should we use for dumpmag? (u_int32_t ?)
> |
> | Since it is used to hold a 32bit unsigned magic number, I'd suggest using
> | u_int32_t everywhere.
> 
> Yeah, I suppose that's the best solution, since changing the constant to be
> 64-bits will break 32-bit kernels.

Ok, I've changed dumpmag to u_int32_t on all ports.

> The alpha port also has this bug and should be fixed at the same time.

On alpha, the first four bytes of u_long is lower four bytes,
so savecore(8) would read (supporsed) correct "32bit" magic number..
---
Izumi Tsutsui
tsutsui@ceres.dti.ne.jp