KVM: propagate fault r/w information to gup(), allow read-only memory
As suggested by Andrea, pass r/w error code to gup(), upgrading read fault
to writable if host pte allows it.
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
diff --git a/arch/x86/kvm/mmu.c b/arch/x86/kvm/mmu.c
index 9943394..53509f5 100644
--- a/arch/x86/kvm/mmu.c
+++ b/arch/x86/kvm/mmu.c
@@ -2216,7 +2216,7 @@
}
static int __direct_map(struct kvm_vcpu *vcpu, gpa_t v, int write,
- int level, gfn_t gfn, pfn_t pfn)
+ int map_writable, int level, gfn_t gfn, pfn_t pfn)
{
struct kvm_shadow_walk_iterator iterator;
struct kvm_mmu_page *sp;
@@ -2225,9 +2225,13 @@
for_each_shadow_entry(vcpu, (u64)gfn << PAGE_SHIFT, iterator) {
if (iterator.level == level) {
- mmu_set_spte(vcpu, iterator.sptep, ACC_ALL, ACC_ALL,
+ unsigned pte_access = ACC_ALL;
+
+ if (!map_writable)
+ pte_access &= ~ACC_WRITE_MASK;
+ mmu_set_spte(vcpu, iterator.sptep, ACC_ALL, pte_access,
0, write, 1, &pt_write,
- level, gfn, pfn, false, true);
+ level, gfn, pfn, false, map_writable);
direct_pte_prefetch(vcpu, iterator.sptep);
++vcpu->stat.pf_fixed;
break;
@@ -2288,6 +2292,7 @@
int level;
pfn_t pfn;
unsigned long mmu_seq;
+ bool map_writable;
level = mapping_level(vcpu, gfn);
@@ -2302,7 +2307,7 @@
mmu_seq = vcpu->kvm->mmu_notifier_seq;
smp_rmb();
- pfn = gfn_to_pfn(vcpu->kvm, gfn);
+ pfn = gfn_to_pfn_prot(vcpu->kvm, gfn, write, &map_writable);
/* mmio */
if (is_error_pfn(pfn))
@@ -2312,7 +2317,7 @@
if (mmu_notifier_retry(vcpu, mmu_seq))
goto out_unlock;
kvm_mmu_free_some_pages(vcpu);
- r = __direct_map(vcpu, v, write, level, gfn, pfn);
+ r = __direct_map(vcpu, v, write, map_writable, level, gfn, pfn);
spin_unlock(&vcpu->kvm->mmu_lock);
@@ -2611,11 +2616,11 @@
}
static bool try_async_pf(struct kvm_vcpu *vcpu, bool no_apf, gfn_t gfn,
- gva_t gva, pfn_t *pfn)
+ gva_t gva, pfn_t *pfn, bool write, bool *writable)
{
bool async;
- *pfn = gfn_to_pfn_async(vcpu->kvm, gfn, &async);
+ *pfn = gfn_to_pfn_async(vcpu->kvm, gfn, &async, write, writable);
if (!async)
return false; /* *pfn has correct page already */
@@ -2632,7 +2637,7 @@
return true;
}
- *pfn = gfn_to_pfn(vcpu->kvm, gfn);
+ *pfn = gfn_to_pfn_prot(vcpu->kvm, gfn, write, writable);
return false;
}
@@ -2645,6 +2650,8 @@
int level;
gfn_t gfn = gpa >> PAGE_SHIFT;
unsigned long mmu_seq;
+ int write = error_code & PFERR_WRITE_MASK;
+ bool map_writable;
ASSERT(vcpu);
ASSERT(VALID_PAGE(vcpu->arch.mmu.root_hpa));
@@ -2660,7 +2667,7 @@
mmu_seq = vcpu->kvm->mmu_notifier_seq;
smp_rmb();
- if (try_async_pf(vcpu, no_apf, gfn, gpa, &pfn))
+ if (try_async_pf(vcpu, no_apf, gfn, gpa, &pfn, write, &map_writable))
return 0;
/* mmio */
@@ -2670,7 +2677,7 @@
if (mmu_notifier_retry(vcpu, mmu_seq))
goto out_unlock;
kvm_mmu_free_some_pages(vcpu);
- r = __direct_map(vcpu, gpa, error_code & PFERR_WRITE_MASK,
+ r = __direct_map(vcpu, gpa, write, map_writable,
level, gfn, pfn);
spin_unlock(&vcpu->kvm->mmu_lock);
diff --git a/arch/x86/kvm/paging_tmpl.h b/arch/x86/kvm/paging_tmpl.h
index d6b281e..ba00eef 100644
--- a/arch/x86/kvm/paging_tmpl.h
+++ b/arch/x86/kvm/paging_tmpl.h
@@ -427,7 +427,7 @@
static u64 *FNAME(fetch)(struct kvm_vcpu *vcpu, gva_t addr,
struct guest_walker *gw,
int user_fault, int write_fault, int hlevel,
- int *ptwrite, pfn_t pfn)
+ int *ptwrite, pfn_t pfn, bool map_writable)
{
unsigned access = gw->pt_access;
struct kvm_mmu_page *sp = NULL;
@@ -501,7 +501,7 @@
mmu_set_spte(vcpu, it.sptep, access, gw->pte_access & access,
user_fault, write_fault, dirty, ptwrite, it.level,
- gw->gfn, pfn, false, true);
+ gw->gfn, pfn, false, map_writable);
FNAME(pte_prefetch)(vcpu, gw, it.sptep);
return it.sptep;
@@ -539,6 +539,7 @@
pfn_t pfn;
int level = PT_PAGE_TABLE_LEVEL;
unsigned long mmu_seq;
+ bool map_writable;
pgprintk("%s: addr %lx err %x\n", __func__, addr, error_code);
@@ -569,13 +570,17 @@
mmu_seq = vcpu->kvm->mmu_notifier_seq;
smp_rmb();
- if (try_async_pf(vcpu, no_apf, walker.gfn, addr, &pfn))
+ if (try_async_pf(vcpu, no_apf, walker.gfn, addr, &pfn, write_fault,
+ &map_writable))
return 0;
/* mmio */
if (is_error_pfn(pfn))
return kvm_handle_bad_page(vcpu->kvm, walker.gfn, pfn);
+ if (!map_writable)
+ walker.pte_access &= ~ACC_WRITE_MASK;
+
spin_lock(&vcpu->kvm->mmu_lock);
if (mmu_notifier_retry(vcpu, mmu_seq))
goto out_unlock;
@@ -583,7 +588,7 @@
trace_kvm_mmu_audit(vcpu, AUDIT_PRE_PAGE_FAULT);
kvm_mmu_free_some_pages(vcpu);
sptep = FNAME(fetch)(vcpu, addr, &walker, user_fault, write_fault,
- level, &write_pt, pfn);
+ level, &write_pt, pfn, map_writable);
(void)sptep;
pgprintk("%s: shadow pte %p %llx ptwrite %d\n", __func__,
sptep, *sptep, write_pt);