#include "vmx.h"
#include "kvm.h"
-#define pgprintk(x...) do { } while (0)
-#define rmap_printk(x...) do { } while (0)
+#define pgprintk(x...) do { printk(x); } while (0)
+#define rmap_printk(x...) do { printk(x); } while (0)
#define ASSERT(x) \
if (!(x)) { \
__FILE__, __LINE__, #x); \
}
-#define PT64_ENT_PER_PAGE 512
-#define PT32_ENT_PER_PAGE 1024
+#define PT64_PT_BITS 9
+#define PT64_ENT_PER_PAGE (1 << PT64_PT_BITS)
+#define PT32_PT_BITS 10
+#define PT32_ENT_PER_PAGE (1 << PT32_PT_BITS)
#define PT_WRITABLE_SHIFT 1
}
}
+static void rmap_write_protect(struct kvm *kvm, u64 gfn)
+{
+ struct page *page;
+ struct kvm_memory_slot *slot;
+ struct kvm_rmap_desc *desc;
+ u64 *spte;
+
+ slot = gfn_to_memslot(kvm, gfn);
+ BUG_ON(!slot);
+ page = gfn_to_page(slot, gfn);
+
+ while (page->private) {
+ if (!(page->private & 1))
+ spte = (u64 *)page->private;
+ else {
+ desc = (struct kvm_rmap_desc *)(page->private & ~1ul);
+ spte = desc->shadow_ptes[0];
+ }
+ BUG_ON(!spte);
+ BUG_ON((*spte & PT64_BASE_ADDR_MASK) !=
+ page_to_pfn(page) << PAGE_SHIFT);
+ BUG_ON(!(*spte & PT_PRESENT_MASK));
+ BUG_ON(!(*spte & PT_WRITABLE_MASK));
+ rmap_printk("rmap_write_protect: spte %p %llx\n", spte, *spte);
+ rmap_remove(kvm, spte);
+ *spte &= ~(u64)PT_WRITABLE_MASK;
+ }
+}
+
static void kvm_mmu_free_page(struct kvm_vcpu *vcpu, hpa_t page_hpa)
{
struct kvm_mmu_page *page_head = page_header(page_hpa);
return 1;
}
-static hpa_t kvm_mmu_alloc_page(struct kvm_vcpu *vcpu, u64 *parent_pte)
+static unsigned kvm_page_table_hashfn(gfn_t gfn)
+{
+ return gfn;
+}
+
+static struct kvm_mmu_page *kvm_mmu_alloc_page(struct kvm_vcpu *vcpu,
+ u64 *parent_pte)
{
struct kvm_mmu_page *page;
if (list_empty(&vcpu->free_pages))
- return INVALID_PAGE;
+ return NULL;
page = list_entry(vcpu->free_pages.next, struct kvm_mmu_page, link);
list_del(&page->link);
ASSERT(is_empty_shadow_page(page->page_hpa));
page->slot_bitmap = 0;
page->global = 1;
+ page->multimapped = 0;
page->parent_pte = parent_pte;
- return page->page_hpa;
+ return page;
+}
+
+static void mmu_page_add_parent_pte(struct kvm_mmu_page *page, u64 *parent_pte)
+{
+ struct kvm_pte_chain *pte_chain;
+ struct hlist_node *node;
+ int i;
+
+ if (!parent_pte)
+ return;
+ if (!page->multimapped) {
+ u64 *old = page->parent_pte;
+
+ if (!old) {
+ page->parent_pte = parent_pte;
+ return;
+ }
+ page->multimapped = 1;
+ pte_chain = kzalloc(sizeof(struct kvm_pte_chain), GFP_NOWAIT);
+ BUG_ON(!pte_chain);
+ INIT_HLIST_HEAD(&page->parent_ptes);
+ hlist_add_head(&pte_chain->link, &page->parent_ptes);
+ pte_chain->parent_ptes[0] = old;
+ }
+ hlist_for_each_entry(pte_chain, node, &page->parent_ptes, link) {
+ if (pte_chain->parent_ptes[NR_PTE_CHAIN_ENTRIES-1])
+ continue;
+ for (i = 0; i < NR_PTE_CHAIN_ENTRIES; ++i)
+ if (!pte_chain->parent_ptes[i]) {
+ pte_chain->parent_ptes[i] = parent_pte;
+ return;
+ }
+ }
+ pte_chain = kzalloc(sizeof(struct kvm_pte_chain), GFP_NOWAIT);
+ BUG_ON(!pte_chain);
+ hlist_add_head(&pte_chain->link, &page->parent_ptes);
+ pte_chain->parent_ptes[0] = parent_pte;
+}
+
+static void mmu_page_remove_parent_pte(struct kvm_mmu_page *page,
+ u64 *parent_pte)
+{
+ struct kvm_pte_chain *pte_chain;
+ struct hlist_node *node;
+ int i;
+
+ if (!page->multimapped) {
+ BUG_ON(page->parent_pte != parent_pte);
+ page->parent_pte = NULL;
+ return;
+ }
+ hlist_for_each_entry(pte_chain, node, &page->parent_ptes, link)
+ for (i = 0; i < NR_PTE_CHAIN_ENTRIES; ++i) {
+ if (!pte_chain->parent_ptes[i])
+ break;
+ if (pte_chain->parent_ptes[i] != parent_pte)
+ continue;
+ while (i + 1 < NR_PTE_CHAIN_ENTRIES) {
+ pte_chain->parent_ptes[i]
+ = pte_chain->parent_ptes[i + 1];
+ ++i;
+ }
+ pte_chain->parent_ptes[i] = NULL;
+ return;
+ }
+ BUG();
+}
+
+static struct kvm_mmu_page *kvm_mmu_lookup_page(struct kvm_vcpu *vcpu,
+ gfn_t gfn)
+{
+ unsigned index;
+ struct hlist_head *bucket;
+ struct kvm_mmu_page *page;
+ struct hlist_node *node;
+
+ pgprintk("%s: looking for gfn %lx\n", __FUNCTION__, gfn);
+ index = kvm_page_table_hashfn(gfn) % KVM_NUM_MMU_PAGES;
+ bucket = &vcpu->kvm->mmu_page_hash[index];
+ hlist_for_each_entry(page, node, bucket, hash_link)
+ if (page->gfn == gfn && !page->role.metaphysical) {
+ pgprintk("%s: found role %x\n",
+ __FUNCTION__, page->role.word);
+ return page;
+ }
+ return NULL;
+}
+
+static struct kvm_mmu_page *kvm_mmu_get_page(struct kvm_vcpu *vcpu,
+ gfn_t gfn,
+ gva_t gaddr,
+ unsigned level,
+ int metaphysical,
+ u64 *parent_pte)
+{
+ union kvm_mmu_page_role role;
+ unsigned index;
+ unsigned quadrant;
+ struct hlist_head *bucket;
+ struct kvm_mmu_page *page;
+ struct hlist_node *node;
+
+ role.word = 0;
+ role.glevels = vcpu->mmu.root_level;
+ role.level = level;
+ role.metaphysical = metaphysical;
+ if (vcpu->mmu.root_level <= PT32_ROOT_LEVEL) {
+ quadrant = gaddr >> (PAGE_SHIFT + (PT64_PT_BITS * level));
+ quadrant &= (1 << ((PT32_PT_BITS - PT64_PT_BITS) * level)) - 1;
+ role.quadrant = quadrant;
+ }
+ pgprintk("%s: looking gfn %lx role %x\n", __FUNCTION__,
+ gfn, role.word);
+ index = kvm_page_table_hashfn(gfn) % KVM_NUM_MMU_PAGES;
+ bucket = &vcpu->kvm->mmu_page_hash[index];
+ hlist_for_each_entry(page, node, bucket, hash_link)
+ if (page->gfn == gfn && page->role.word == role.word) {
+ mmu_page_add_parent_pte(page, parent_pte);
+ pgprintk("%s: found\n", __FUNCTION__);
+ return page;
+ }
+ page = kvm_mmu_alloc_page(vcpu, parent_pte);
+ if (!page)
+ return page;
+ pgprintk("%s: adding gfn %lx role %x\n", __FUNCTION__, gfn, role.word);
+ page->gfn = gfn;
+ page->role = role;
+ hlist_add_head(&page->hash_link, bucket);
+ if (!metaphysical)
+ rmap_write_protect(vcpu->kvm, gfn);
+ return page;
+}
+
+static void kvm_mmu_put_page(struct kvm_vcpu *vcpu,
+ struct kvm_mmu_page *page,
+ u64 *parent_pte)
+{
+ mmu_page_remove_parent_pte(page, parent_pte);
}
static void page_header_update_slot(struct kvm *kvm, void *pte, gpa_t gpa)
for (; ; level--) {
u32 index = PT64_INDEX(v, level);
u64 *table;
+ u64 pte;
ASSERT(VALID_PAGE(table_addr));
table = __va(table_addr);
if (level == 1) {
+ pte = table[index];
+ if (is_present_pte(pte) && is_writeble_pte(pte))
+ return 0;
mark_page_dirty(vcpu->kvm, v >> PAGE_SHIFT);
page_header_update_slot(vcpu->kvm, table, v);
table[index] = p | PT_PRESENT_MASK | PT_WRITABLE_MASK |
}
if (table[index] == 0) {
- hpa_t new_table = kvm_mmu_alloc_page(vcpu,
- &table[index]);
-
- if (!VALID_PAGE(new_table)) {
+ struct kvm_mmu_page *new_table;
+ gfn_t pseudo_gfn;
+
+ pseudo_gfn = (v & PT64_DIR_BASE_ADDR_MASK)
+ >> PAGE_SHIFT;
+ new_table = kvm_mmu_get_page(vcpu, pseudo_gfn,
+ v, level - 1,
+ 1, &table[index]);
+ if (!new_table) {
pgprintk("nonpaging_map: ENOMEM\n");
return -ENOMEM;
}
- if (level == PT32E_ROOT_LEVEL)
- table[index] = new_table | PT_PRESENT_MASK;
- else
- table[index] = new_table | PT_PRESENT_MASK |
- PT_WRITABLE_MASK | PT_USER_MASK;
+ table[index] = new_table->page_hpa | PT_PRESENT_MASK
+ | PT_WRITABLE_MASK | PT_USER_MASK;
}
table_addr = table[index] & PT64_BASE_ADDR_MASK;
}
hpa_t root = vcpu->mmu.root_hpa;
ASSERT(VALID_PAGE(root));
- release_pt_page_64(vcpu, root, PT64_ROOT_LEVEL);
vcpu->mmu.root_hpa = INVALID_PAGE;
return;
}
ASSERT(VALID_PAGE(root));
root &= PT64_BASE_ADDR_MASK;
- release_pt_page_64(vcpu, root, PT32E_ROOT_LEVEL - 1);
vcpu->mmu.pae_root[i] = INVALID_PAGE;
}
vcpu->mmu.root_hpa = INVALID_PAGE;
static void mmu_alloc_roots(struct kvm_vcpu *vcpu)
{
int i;
+ gfn_t root_gfn;
+ root_gfn = vcpu->cr3 >> PAGE_SHIFT;
#ifdef CONFIG_X86_64
if (vcpu->mmu.shadow_root_level == PT64_ROOT_LEVEL) {
hpa_t root = vcpu->mmu.root_hpa;
ASSERT(!VALID_PAGE(root));
- root = kvm_mmu_alloc_page(vcpu, NULL);
+ root = kvm_mmu_get_page(vcpu, root_gfn, 0,
+ PT64_ROOT_LEVEL, 0, NULL)->page_hpa;
vcpu->mmu.root_hpa = root;
return;
}
hpa_t root = vcpu->mmu.pae_root[i];
ASSERT(!VALID_PAGE(root));
- root = kvm_mmu_alloc_page(vcpu, NULL);
+ if (vcpu->mmu.root_level == PT32E_ROOT_LEVEL)
+ root_gfn = vcpu->pdptrs[i] >> PAGE_SHIFT;
+ else if (vcpu->mmu.root_level == 0)
+ root_gfn = 0;
+ root = kvm_mmu_get_page(vcpu, root_gfn, i << 30,
+ PT32_ROOT_LEVEL, !is_paging(vcpu),
+ NULL)->page_hpa;
vcpu->mmu.pae_root[i] = root | PT_PRESENT_MASK;
}
vcpu->mmu.root_hpa = __pa(vcpu->mmu.pae_root);
context->inval_page = nonpaging_inval_page;
context->gva_to_gpa = nonpaging_gva_to_gpa;
context->free = nonpaging_free;
- context->root_level = PT32E_ROOT_LEVEL;
+ context->root_level = 0;
context->shadow_root_level = PT32E_ROOT_LEVEL;
mmu_alloc_roots(vcpu);
ASSERT(VALID_PAGE(context->root_hpa));
return 0;
}
-
static void kvm_mmu_flush_tlb(struct kvm_vcpu *vcpu)
{
- struct kvm_mmu_page *page, *npage;
-
- list_for_each_entry_safe(page, npage, &vcpu->kvm->active_mmu_pages,
- link) {
- if (page->global)
- continue;
-
- if (!page->parent_pte)
- continue;
-
- *page->parent_pte = 0;
- release_pt_page_64(vcpu, page->page_hpa, 1);
- }
++kvm_stat.tlb_flush;
kvm_arch_ops->tlb_flush(vcpu);
}
static void paging_new_cr3(struct kvm_vcpu *vcpu)
{
+ pgprintk("%s: cr3 %lx\n", __FUNCTION__, vcpu->cr3);
+ mmu_free_roots(vcpu);
+ mmu_alloc_roots(vcpu);
kvm_mmu_flush_tlb(vcpu);
+ kvm_arch_ops->set_cr3(vcpu, vcpu->mmu.root_hpa);
}
static void mark_pagetable_nonglobal(void *shadow_pte)
if (!dirty)
access_bits &= ~PT_WRITABLE_MASK;
- if (access_bits & PT_WRITABLE_MASK)
- mark_page_dirty(vcpu->kvm, gaddr >> PAGE_SHIFT);
+ paddr = gpa_to_hpa(vcpu, gaddr & PT64_BASE_ADDR_MASK);
*shadow_pte |= access_bits;
- paddr = gpa_to_hpa(vcpu, gaddr & PT64_BASE_ADDR_MASK);
-
if (!(*shadow_pte & PT_GLOBAL_MASK))
mark_pagetable_nonglobal(shadow_pte);
*shadow_pte |= gaddr;
*shadow_pte |= PT_SHADOW_IO_MARK;
*shadow_pte &= ~PT_PRESENT_MASK;
- } else {
- *shadow_pte |= paddr;
- page_header_update_slot(vcpu->kvm, shadow_pte, gaddr);
- rmap_add(vcpu->kvm, shadow_pte);
+ return;
}
+
+ *shadow_pte |= paddr;
+
+ if (access_bits & PT_WRITABLE_MASK) {
+ struct kvm_mmu_page *shadow;
+
+ shadow = kvm_mmu_lookup_page(vcpu, gaddr >> PAGE_SHIFT);
+ if (shadow) {
+ pgprintk("%s: found shadow page for %lx, marking ro\n",
+ __FUNCTION__, (gfn_t)(gaddr >> PAGE_SHIFT));
+ access_bits &= ~PT_WRITABLE_MASK;
+ *shadow_pte &= ~PT_WRITABLE_MASK;
+ }
+ }
+
+ if (access_bits & PT_WRITABLE_MASK)
+ mark_page_dirty(vcpu->kvm, gaddr >> PAGE_SHIFT);
+
+ page_header_update_slot(vcpu->kvm, shadow_pte, gaddr);
+ rmap_add(vcpu->kvm, shadow_pte);
}
static void inject_page_fault(struct kvm_vcpu *vcpu,
projects / linux-block.git / blobdiff