[PATCH] KVM: MMU: Implement child shadow unlinking

When removing a page table, we must maintain the parent_pte field all child
shadow page tables.

Signed-off-by: Avi Kivity <avi@qumranet.com>
Acked-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

diff --git a/drivers/kvm/mmu.c b/drivers/kvm/mmu.c
index 1484b721171718c645d3d5eee85f8585cd745969..7e20dbf4f84cf7e1644f8234a287ac3f6fb7627f 100644 (file)
--- a/drivers/kvm/mmu.c
+++ b/drivers/kvm/mmu.c
@@ -402,12 +402,21 @@ static void mmu_page_remove_parent_pte(struct kvm_mmu_page *page,
                                break;
                        if (pte_chain->parent_ptes[i] != parent_pte)
                                continue;
-                       while (i + 1 < NR_PTE_CHAIN_ENTRIES) {
+                       while (i + 1 < NR_PTE_CHAIN_ENTRIES
+                               && pte_chain->parent_ptes[i + 1]) {
                                pte_chain->parent_ptes[i]
                                        = pte_chain->parent_ptes[i + 1];
                                ++i;
                        }
                        pte_chain->parent_ptes[i] = NULL;
+                       if (i == 0) {
+                               hlist_del(&pte_chain->link);
+                               kfree(pte_chain);
+                               if (hlist_empty(&page->parent_ptes)) {
+                                       page->multimapped = 0;
+                                       page->parent_pte = NULL;
+                               }
+                       }
                        return;
                }
        BUG();
@@ -481,7 +490,30 @@ static struct kvm_mmu_page *kvm_mmu_get_page(struct kvm_vcpu *vcpu,
 static void kvm_mmu_page_unlink_children(struct kvm_vcpu *vcpu,
                                         struct kvm_mmu_page *page)
 {
-       BUG();
+       unsigned i;
+       u64 *pt;
+       u64 ent;
+
+       pt = __va(page->page_hpa);
+
+       if (page->role.level == PT_PAGE_TABLE_LEVEL) {
+               for (i = 0; i < PT64_ENT_PER_PAGE; ++i) {
+                       if (pt[i] & PT_PRESENT_MASK)
+                               rmap_remove(vcpu->kvm, &pt[i]);
+                       pt[i] = 0;
+               }
+               return;
+       }
+
+       for (i = 0; i < PT64_ENT_PER_PAGE; ++i) {
+               ent = pt[i];
+
+               pt[i] = 0;
+               if (!(ent & PT_PRESENT_MASK))
+                       continue;
+               ent &= PT64_BASE_ADDR_MASK;
+               mmu_page_remove_parent_pte(page_header(ent), &pt[i]);
+       }
 }
 
 static void kvm_mmu_put_page(struct kvm_vcpu *vcpu,
@@ -489,8 +521,7 @@ static void kvm_mmu_put_page(struct kvm_vcpu *vcpu,
                             u64 *parent_pte)
 {
        mmu_page_remove_parent_pte(page, parent_pte);
-       if (page->role.level > PT_PAGE_TABLE_LEVEL)
-               kvm_mmu_page_unlink_children(vcpu, page);
+       kvm_mmu_page_unlink_children(vcpu, page);
        hlist_del(&page->hash_link);
        list_del(&page->link);
        list_add(&page->link, &vcpu->free_pages);
@@ -511,6 +542,7 @@ static void kvm_mmu_zap_page(struct kvm_vcpu *vcpu,
                                             struct kvm_pte_chain, link);
                        parent_pte = chain->parent_ptes[0];
                }
+               BUG_ON(!parent_pte);
                kvm_mmu_put_page(vcpu, page, parent_pte);
                *parent_pte = 0;
        }
@@ -530,6 +562,8 @@ static int kvm_mmu_unprotect_page(struct kvm_vcpu *vcpu, gfn_t gfn)
        bucket = &vcpu->kvm->mmu_page_hash[index];
        hlist_for_each_entry_safe(page, node, n, bucket, hash_link)
                if (page->gfn == gfn && !page->role.metaphysical) {
+                       pgprintk("%s: gfn %lx role %x\n", __FUNCTION__, gfn,
+                                page->role.word);
                        kvm_mmu_zap_page(vcpu, page);
                        r = 1;
                }