[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index] RE: [PATCH 1/6] VT-d: properly reserve DID 0 for caching mode IOMMUs
> From: Roger Pau Monné <roger.pau@xxxxxxxxxx> > Sent: Friday, November 12, 2021 8:19 PM > > On Fri, Nov 12, 2021 at 01:07:33PM +0100, Jan Beulich wrote: > > On 12.11.2021 12:23, Roger Pau Monné wrote: > > > On Fri, Nov 12, 2021 at 10:47:59AM +0100, Jan Beulich wrote: > > >> Merely setting bit 0 in the bitmap is insufficient, as then Dom0 will > > >> still have DID 0 allocated to it, because of the zero-filling of > > >> domid_map[]. Set slot 0 to DOMID_INVALID to keep DID 0 from getting > > >> used. > > > > > > Shouldn't the whole domid_map be initialized to DOMID_INVALID to > > > prevent dom0 matching against any unused slot? > > > > > > Similarly cleanup_domid_map should set the slot to DOMID_INVALID. > > > > I don't think so, that's the purpose of setting the bit in domid_bitmap. > > The problem really was only with setting a bit in that bitmap without > > invalidating the corresponding slot. > > > > This said, I can still see value in doing as you suggest, but as a > > separate change with a different justification. In fact domid_bitmap is > > kind of redundant now anyway; aiui it was the thing that existed first. > > Then domid_map[] was simply added, rather than fully replacing the > > original bitmap. > > I guess using domid_bitmap to find a free slot is faster than scanning > the array of iommu IDs to domids. Not sure how performance critical > that search is, so maybe it's fine to just drop domid_bitmap and rely > exclusively on the array. > I'm fine to drop domid_bitmap. I don't think domain creation is in hot path...
|
![]() |
Lists.xenproject.org is hosted with RackSpace, monitoring our |