On Tue, Nov 21, 2006 at 02:12:33PM +0200, Constantine Kousoulos wrote: > Thomas Bushnell BSG wrote: > >On Mon, 2006-11-20 at 16:15 +0200, Constantine Kousoulos wrote: > >>All the above solutions aim to make Mach unaware of any memory > >>above 1 GB. Isn't it possible to utilise all the memory? I will > >>explain myself. > > > >I think this is clearly right, but it may well require more substantial > >changes to actually make it work. > > > > Can you give us a few directions?
It really depends on the way you implement it. Mach is damn slow, let's not make it slower by changing the kernel address space in a way that would prevent it from being included in user address spaces. On the other hand, implementing something like kmap() is, as Samuel said, bug-prone. And again, I really don't think changing such low level stuff in GNU Mach (which would automatically create new bugs) is a priority, since 512/1024 MB are very common RAM sizes. In other words, I'd suggest staying with the suggested changes to limit the amount of physical memory to 1 GB until someone *really* takes the time to propose and implement a suitable solution for handling more memory. -- Richard Braun
signature.asc
Description: Digital signature
_______________________________________________ Bug-hurd mailing list Bug-hurd@gnu.org http://lists.gnu.org/mailman/listinfo/bug-hurd
