On 17.08.2021 19:25, Andrew Cooper wrote: > On 17/08/2021 15:29, Jan Beulich wrote: >> About every time I look at dom0_construct_pv()'s "calculation" of >> nr_pt_pages I question (myself) whether the result is precise or merely >> an upper bound. I think it is meant to be precise, but I think we would >> be better off having some checking in place. Hence add ASSERT()s to >> verify that >> - all pages have a valid L1...Ln (currently L4) page table type and >> - no other bits are set, in particular the type refcount is still zero. >> >> Signed-off-by: Jan Beulich <[email protected]> > > Reviewed-by: Andrew Cooper <[email protected]>, thanks. > > Are you planning further cleanups here imminently? If not, I can > probably shuffle some of the easy ROUNDUP() refactoring in the direction > of an intern or grad.
Not any cleanup, I don't think, but quite a few further changes to make Dom0 use large IOMMU page mappings where possible (without introducing logic yet to un-shatter split pages, not the least because relying on just that would then undermine the secondary effect of improving boot time). The two changes posted so far were really just fallout from that work, with it seeming reasonable to post up front to reduce the size of the actual series, which has grown quite a bit longer than I though it would. Jan
