I have just pushed out Syslinux 3.80-pre7. I am slightly concerned
about putting this into the release tree at this point: I was working
with some of the shuffler code, and found that the memory map handling
was actually severely buggy on some of my test cases. Resolving that
problem took rewriting most of syslinux_add_memmap().
However, against better judgement :) I decided to check in the other
work I have been doing as well (make the shuffler correctly handle cases
where as single source region is supposed to be copied to more than one
destination region.) That is probably a bit aggressive, but I'm
justifying it with the fact that that case was already buggy, and that
if there are no such cases in the shuffle map, the code should be a
no-op as it doesn't affect the inner loop of the shuffler(*).
Hence, any testing support would be appreciated...
-hpa
(*) The algorithm used is based on the observation that any combination
of one-to-many mappings can be treated as a one-to-one shuffle followed
by a set of destination-to-destination copies...
--
H. Peter Anvin, Intel Open Source Technology Center
I work for Intel. I don't speak on their behalf.