KVM: MMU: add SPTE_HOST_WRITEABLE flag to the shadow ptes
Izik Eidus [Wed, 23 Sep 2009 18:47:17 +0000 (21:47 +0300)]
this flag notify that the host physical page we are pointing to from
the spte is write protected, and therefore we cant change its access
to be write unless we run get_user_pages(write = 1).

(this is needed for change_pte support in kvm)

Signed-off-by: Izik Eidus <ieidus@redhat.com>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>

arch/x86/kvm/mmu.c
arch/x86/kvm/paging_tmpl.h

index 6c67b23..5cd8b4e 100644 (file)
@@ -156,6 +156,8 @@ module_param(oos_shadow, bool, 0644);
 #define CREATE_TRACE_POINTS
 #include "mmutrace.h"
 
+#define SPTE_HOST_WRITEABLE (1ULL << PT_FIRST_AVAIL_BITS_SHIFT)
+
 #define SHADOW_PT_INDEX(addr, level) PT64_INDEX(addr, level)
 
 struct kvm_rmap_desc {
@@ -1754,7 +1756,7 @@ static int set_spte(struct kvm_vcpu *vcpu, u64 *sptep,
                    unsigned pte_access, int user_fault,
                    int write_fault, int dirty, int level,
                    gfn_t gfn, pfn_t pfn, bool speculative,
-                   bool can_unsync)
+                   bool can_unsync, bool reset_host_protection)
 {
        u64 spte;
        int ret = 0;
@@ -1781,6 +1783,9 @@ static int set_spte(struct kvm_vcpu *vcpu, u64 *sptep,
                spte |= kvm_x86_ops->get_mt_mask(vcpu, gfn,
                        kvm_is_mmio_pfn(pfn));
 
+       if (reset_host_protection)
+               spte |= SPTE_HOST_WRITEABLE;
+
        spte |= (u64)pfn << PAGE_SHIFT;
 
        if ((pte_access & ACC_WRITE_MASK)
@@ -1826,7 +1831,8 @@ static void mmu_set_spte(struct kvm_vcpu *vcpu, u64 *sptep,
                         unsigned pt_access, unsigned pte_access,
                         int user_fault, int write_fault, int dirty,
                         int *ptwrite, int level, gfn_t gfn,
-                        pfn_t pfn, bool speculative)
+                        pfn_t pfn, bool speculative,
+                        bool reset_host_protection)
 {
        int was_rmapped = 0;
        int was_writeble = is_writeble_pte(*sptep);
@@ -1858,7 +1864,8 @@ static void mmu_set_spte(struct kvm_vcpu *vcpu, u64 *sptep,
        }
 
        if (set_spte(vcpu, sptep, pte_access, user_fault, write_fault,
-                     dirty, level, gfn, pfn, speculative, true)) {
+                     dirty, level, gfn, pfn, speculative, true,
+                     reset_host_protection)) {
                if (write_fault)
                        *ptwrite = 1;
                kvm_x86_ops->tlb_flush(vcpu);
@@ -1906,7 +1913,7 @@ static int __direct_map(struct kvm_vcpu *vcpu, gpa_t v, int write,
                if (iterator.level == level) {
                        mmu_set_spte(vcpu, iterator.sptep, ACC_ALL, ACC_ALL,
                                     0, write, 1, &pt_write,
-                                    level, gfn, pfn, false);
+                                    level, gfn, pfn, false, true);
                        ++vcpu->stat.pf_fixed;
                        break;
                }
index d2fec9c..72558f8 100644 (file)
@@ -273,9 +273,13 @@ static void FNAME(update_pte)(struct kvm_vcpu *vcpu, struct kvm_mmu_page *page,
        if (mmu_notifier_retry(vcpu, vcpu->arch.update_pte.mmu_seq))
                return;
        kvm_get_pfn(pfn);
+       /*
+        * we call mmu_set_spte() with reset_host_protection = true beacuse that
+        * vcpu->arch.update_pte.pfn was fetched from get_user_pages(write = 1).
+        */
        mmu_set_spte(vcpu, spte, page->role.access, pte_access, 0, 0,
                     gpte & PT_DIRTY_MASK, NULL, PT_PAGE_TABLE_LEVEL,
-                    gpte_to_gfn(gpte), pfn, true);
+                    gpte_to_gfn(gpte), pfn, true, true);
 }
 
 /*
@@ -308,7 +312,7 @@ static u64 *FNAME(fetch)(struct kvm_vcpu *vcpu, gva_t addr,
                                     user_fault, write_fault,
                                     gw->ptes[gw->level-1] & PT_DIRTY_MASK,
                                     ptwrite, level,
-                                    gw->gfn, pfn, false);
+                                    gw->gfn, pfn, false, true);
                        break;
                }
 
@@ -558,6 +562,7 @@ static void FNAME(prefetch_page)(struct kvm_vcpu *vcpu,
 static int FNAME(sync_page)(struct kvm_vcpu *vcpu, struct kvm_mmu_page *sp)
 {
        int i, offset, nr_present;
+       bool reset_host_protection;
 
        offset = nr_present = 0;
 
@@ -595,9 +600,16 @@ static int FNAME(sync_page)(struct kvm_vcpu *vcpu, struct kvm_mmu_page *sp)
 
                nr_present++;
                pte_access = sp->role.access & FNAME(gpte_access)(vcpu, gpte);
+               if (!(sp->spt[i] & SPTE_HOST_WRITEABLE)) {
+                       pte_access &= ~ACC_WRITE_MASK;
+                       reset_host_protection = 0;
+               } else {
+                       reset_host_protection = 1;
+               }
                set_spte(vcpu, &sp->spt[i], pte_access, 0, 0,
                         is_dirty_gpte(gpte), PT_PAGE_TABLE_LEVEL, gfn,
-                        spte_to_pfn(sp->spt[i]), true, false);
+                        spte_to_pfn(sp->spt[i]), true, false,
+                        reset_host_protection);
        }
 
        return !nr_present;