On 04/23/2012 07:48 PM, Rik van Riel wrote: > On 04/23/2012 12:40 PM, Ying Han wrote: > >> Avi, does this patch help the case as you mentioned above, where kvm >> module is loaded but no virtual machines are present ? Why we have to >> walk the empty while holding the spinlock? > > It might help marginally, but rather than defending > the patch, it might be more useful to try finding a > solution to the problem Mike and Eric are triggering. > > How can we prevent the code from taking the lock and > throwing out EPT/NPT pages when there is no real need > to, and they will be faulted back in soon anyway? We need to split the lru into an active and inactive list, just like the linux mm. The inactive list might be global (rather than vm local). Then we can just examine the inactive list, see that it's empty, and quit. Perhaps signal the guests to move some pages from the active to inactive list. This is hindered by lack of accessed bits on EPT page tables. That means we must resort to faults to reactivate shadow pages. -- error compiling committee.c: too many arguments to function -- To unsubscribe from this list: send the line "unsubscribe kvm" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
