[kvm-devel] [1/3] -reserved-ram for PCI passthrough without VT-d and without paravirt

[Andrea A. <an...@qu...>]

This is the kvm.git patch to enable -reserved-ram (without this kvm
will simply gracefully fail to emulate the illegal instruction inside
the bad_page). This trick avoids altering the ioctl api with libkvm,
in short if get_user_pages fails on a host kernel with reserved ram
config option enabled, it tries to see if it's a remap_pfn_range
mapping backing the memslot. In such case it checks if the ram was
reserved with page_count 0 and if so it disables the reference
counting as those pages are invisibile to linux. As long as pfn_valid
returns, pfn_to_page should be safe, so shall the memmap array be
allocated with holes corresponding to the holes generated in the e820
map, simply bad_page will be returned gracefully without risk like if
this patch wasn't applied to kvm.git.

Signed-off-by: Andrea Arcangeli <an...@qu...>

diff --git a/arch/x86/kvm/paging_tmpl.h b/arch/x86/kvm/paging_tmpl.h
index e9ae5db..e7a9c82 100644
--- a/arch/x86/kvm/paging_tmpl.h
+++ b/arch/x86/kvm/paging_tmpl.h
@@ -263,7 +263,8 @@ static void FNAME(update_pte)(struct kvm_vcpu *vcpu, struct kvm_mmu_page *page,
 	npage = vcpu->arch.update_pte.page;
 	if (!npage)
 		return;
-	get_page(npage);
+	if (!page_is_reserved(npage))
+		get_page(npage);
 	mmu_set_spte(vcpu, spte, page->role.access, pte_access, 0, 0,
 		     gpte & PT_DIRTY_MASK, NULL, largepage, gpte_to_gfn(gpte),
 		     npage, true);
diff --git a/include/linux/kvm_host.h b/include/linux/kvm_host.h
index f4e1436..7f087ac 100644
--- a/include/linux/kvm_host.h
+++ b/include/linux/kvm_host.h
@@ -281,6 +281,18 @@ static inline void kvm_migrate_apic_timer(struct kvm_vcpu *vcpu)
 	set_bit(KVM_REQ_MIGRATE_TIMER, &vcpu->requests);
 }
 
+#ifdef CONFIG_RESERVE_PHYSICAL_START
+static inline int page_is_reserved(struct page * page)
+{
+	return !page_count(page);
+}
+#else /* CONFIG_RESERVE_PHYSICAL_START */
+static inline int page_is_reserved(struct page * page)
+{
+	return 0;
+}
+#endif /* CONFIG_RESERVE_PHYSICAL_START */
+
 enum kvm_stat_kind {
 	KVM_STAT_VM,
 	KVM_STAT_VCPU,
diff --git a/virt/kvm/kvm_main.c b/virt/kvm/kvm_main.c
index 30bf832..50a7b3e 100644
--- a/virt/kvm/kvm_main.c
+++ b/virt/kvm/kvm_main.c
@@ -498,6 +524,65 @@ unsigned long gfn_to_hva(struct kvm *kvm, gfn_t gfn)
 	return (slot->userspace_addr + (gfn - slot->base_gfn) * PAGE_SIZE);
 }
 
+
+#ifdef CONFIG_RESERVE_PHYSICAL_START
+static struct page *direct_page(struct mm_struct *mm,
+				unsigned long address)
+{
+	pgd_t *pgd;
+	pud_t *pud;
+	pmd_t *pmd;
+	pte_t *ptep, pte;
+	spinlock_t *ptl;
+	struct page *page;
+	struct vm_area_struct *vma;
+	unsigned long pfn;
+
+	page = NULL;
+	if (!capable(CAP_SYS_ADMIN)) /* go safe */
+		goto out;
+
+	vma = find_vma(current->mm, address);
+	if (!vma || vma->vm_start > address ||
+	    !(vma->vm_flags & VM_PFNMAP))
+		goto out;
+
+	pgd = pgd_offset(mm, address);
+	if (pgd_none(*pgd) || unlikely(pgd_bad(*pgd)))
+		goto out;
+
+	pud = pud_offset(pgd, address);
+	if (pud_none(*pud) || unlikely(pud_bad(*pud)))
+		goto out;
+	
+	pmd = pmd_offset(pud, address);
+	if (pmd_none(*pmd) || unlikely(pmd_bad(*pmd)))
+		goto out;
+
+	ptep = pte_offset_map_lock(mm, pmd, address, &ptl);
+	if (!ptep)
+		goto out;
+
+	pte = *ptep;
+	if (!pte_present(pte))
+		goto unlock;
+
+	pfn = pte_pfn(pte);
+	if (!pfn_valid(pfn))
+		goto unlock;
+
+	page = pfn_to_page(pfn);
+	if (!page_is_reserved(page)) {
+		page = NULL;
+		goto unlock;
+	}
+unlock:
+	pte_unmap_unlock(ptep, ptl);
+out:
+	return page;
+}
+#endif /* CONFIG_RESERVE_PHYSICAL_START */
+
 /*
  * Requires current->mm->mmap_sem to be held
  */
@@ -519,6 +604,11 @@ struct page *gfn_to_page(struct kvm *kvm, gfn_t gfn)
 				NULL);
 
 	if (npages != 1) {
+#ifdef CONFIG_RESERVE_PHYSICAL_START
+		page[0] = direct_page(current->mm, addr);
+		if (page[0])
+			return page[0];
+#endif
 		get_page(bad_page);
 		return bad_page;
 	}
@@ -530,15 +620,18 @@ EXPORT_SYMBOL_GPL(gfn_to_page);
 
 void kvm_release_page_clean(struct page *page)
 {
-	put_page(page);
+	if (!page_is_reserved(page))
+		put_page(page);
 }
 EXPORT_SYMBOL_GPL(kvm_release_page_clean);
 
 void kvm_release_page_dirty(struct page *page)
 {
-	if (!PageReserved(page))
-		SetPageDirty(page);
-	put_page(page);
+	if (!page_is_reserved(page)) {
+		if (!PageReserved(page))
+			SetPageDirty(page);
+		put_page(page);
+	}
 }
 EXPORT_SYMBOL_GPL(kvm_release_page_dirty);