>From: Jes Sorensen >Sent: 2006年6月6日 17:15 > >>>>>> "Alex" == Alex Williamson <alex.williamson@xxxxxx> writes: > >Alex> This patch includes only the modifications to existing >Alex> Linux/ia64 files to support both privileged and un-privileged >Alex> guests on Xen/ia64. > >- In general I don't like this CONFIG_XEN being added everywhere. It > would be much nicer if we could seperate the majority of the Xen > code into a seperate file that is only compiled in when enabled. >- In many cases expanding and using the machvec table would reduce >the > ugliness of the patches. For instance in the dma mapping code, > simply define a set of Xen specific functions that do all the magic > stuff nobody else cares about. Hi, Jes, Thanks for careful review. I agree that xen should be introduced as some type of new 'box' as others, like current sn/hp/dig under arch/ia64. That does save many ugly checks for is_running_on_xen scattered in common code. Future it would be even worse. Actually that's the way how xen/x86 has already done currently, where mach-xen is introduced under arch/i386 and thus the hierarchy is much cleaner. IMO, that's one of the outstanding gaps between xen/x86 and xen/ia64, and time for change now. :-) Thanks, Kevin - : send the line "unsubscribe linux-ia64" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
|