On Saturday 13 September 2008 06:11:02 Matias Zabaljauregui wrote: > Hello Rusty, thanks for your replay. > > >> - the launcher for PAE should be a different binary? or could just > >> receive an option from the command line ? (I'm using the command line > >> option --pae). The only difference is the initial page table. > > > > If we can move the page table creation into the kernel, we can avoid > > this distinction altogether. I think that's probably best... > > Ok, I will try this idea. > When you say we can move the page table creation into the kernel, > you mean host code ( we could build the guest initial page table in > function init_guest_pagetable(), included in page_tables.c ) ? > or you refer to guest code (probably arch/x86/lguest/i386_head.S) ?
I think the host will have to do most of it: the guest needs some code mapped if it were to do it itself, and if the host does that, it might as well map the entire memory linear to start. Cheers, Rusty. _______________________________________________ Lguest mailing list [email protected] https://ozlabs.org/mailman/listinfo/lguest
