On 9/19/07, Avi Kivity <[EMAIL PROTECTED]> wrote: > Neo Jia wrote: > > Avi, > > > > Sorry for sending such a direct email to you! > > > > > > I really prefer posts to the list where others can help out. >
Sorry. Just copying this message to list. > > I am thinking to contribute some to KVM project and am very interested > > in changing the mmu page eviction algorithm from FIFO to LRU. > > > > Do you have any schedule in mind? Can you point me roughly to the code? > > > > The entry point is kvm_mmu_free_some_pages() which takes pages off the > lru, however no effort is done to make sure the lru is really in lru > order. I will start from here. This is particularly difficult since shadow page tables are > accessed by the hardware so the code has no chance to change the list > ordering. > > One way to do this is to remove all ptes pointing to a shadow page, so > that the next access will cause a fault and a chance to hoist it back to > the top of the lru. Kind of like the Linux active/inactive lists. > > > Note that others are also looking into this. > > -- > error compiling committee.c: too many arguments to function > > -- I would remember that if researchers were not ambitious probably today we haven't the technology we are using! ------------------------------------------------------------------------- This SF.net email is sponsored by: Microsoft Defy all challenges. Microsoft(R) Visual Studio 2005. http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/ _______________________________________________ kvm-devel mailing list kvm-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/kvm-devel
