[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index] Re: [Xen-devel] [PATCH 2/3] IOMMU/x86: make page type checks consistent when mapping pages
>>> On 13.05.19 at 15:44, <george.dunlap@xxxxxxxxxx> wrote: > On 3/5/19 1:26 PM, Jan Beulich wrote: >> --- a/xen/drivers/passthrough/iommu.c >> +++ b/xen/drivers/passthrough/iommu.c >> @@ -192,21 +192,27 @@ void __hwdom_init iommu_hwdom_init(struc >> >> page_list_for_each ( page, &d->page_list ) >> { >> - unsigned long mfn = mfn_x(page_to_mfn(page)); >> - unsigned long dfn = mfn_to_gmfn(d, mfn); >> - unsigned int mapping = IOMMUF_readable; >> - int ret; >> + if ( (page->u.inuse.type_info & PGT_type_mask) == PGT_none ) >> + { >> + ASSERT(!(page->u.inuse.type_info & PGT_count_mask)); >> + if ( get_page_and_type(page, d, PGT_writable_page) ) >> + put_page_and_type(page); >> + else if ( !rc ) >> + rc = -EBUSY; >> + } >> >> - if ( ((page->u.inuse.type_info & PGT_count_mask) == 0) || >> - ((page->u.inuse.type_info & PGT_type_mask) >> - == PGT_writable_page) ) >> - mapping |= IOMMUF_writable; >> + if ( ((page->u.inuse.type_info & PGT_type_mask) == >> + PGT_writable_page) ) >> + { >> + unsigned long mfn = mfn_x(page_to_mfn(page)); >> + unsigned long dfn = mfn_to_gmfn(d, mfn); >> + int ret = iommu_map(d, _dfn(dfn), _mfn(mfn), 0, >> + IOMMUF_readable | IOMMUF_writable, >> + &flush_flags); > > What's the idea behind calling iommu_map() here, rather than relying on > the one in _get_page_type()? Does need_iommu_pt_sync() not work yet at > this point, or do you expect there to be pages that have been marked > PGT_writable without having gone through _get_page_type()? No, I think I simply didn't realize that this could be deleted altogether with the added get_page_and_type() invocation. Jan _______________________________________________ Xen-devel mailing list Xen-devel@xxxxxxxxxxxxxxxxxxxx https://lists.xenproject.org/mailman/listinfo/xen-devel
|
Lists.xenproject.org is hosted with RackSpace, monitoring our |