On Tue, Mar 24, 2009 at 01:02:10AM -0400, Donald T Hayford wrote:
April 26 won't build. Two lint issues.
April 27 builds and run. 0 errors in 144 succesive MD5 computations.
April 28 builds, but won't run unless you add the following line to
std.nslu2 (before you build the kernel):
options KERNEL_BASE_EXT=0xc0000000
The result is 3 errors in 144 successive MD5 computations.
While it is definite that there is a problem with 20080428 (and all
later kernels), it is less sure that there isn't a problem with
20080427. I think Andy Ruhl was planning to build this version and test
it as well.
http://mail-index.netbsd.org/source-changes/2008/04/27/msg005400.html
That suggests fairly strongly that you may want to try building
earlier kernels from the matt-armv6 branch...