KVM: MMU: optimize to handle dirty bit
If dirty bit is not set, we can make the pte access read-only to avoid handing
dirty bit everywhere
Signed-off-by: Xiao Guangrong <xiaoguangrong@cn.fujitsu.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
diff --git a/arch/x86/kvm/mmu.c b/arch/x86/kvm/mmu.c
index d1986b7..98812c2 100644
--- a/arch/x86/kvm/mmu.c
+++ b/arch/x86/kvm/mmu.c
@@ -1923,7 +1923,7 @@
static int set_spte(struct kvm_vcpu *vcpu, u64 *sptep,
unsigned pte_access, int user_fault,
- int write_fault, int dirty, int level,
+ int write_fault, int level,
gfn_t gfn, pfn_t pfn, bool speculative,
bool can_unsync, bool host_writable)
{
@@ -1938,8 +1938,7 @@
spte = PT_PRESENT_MASK;
if (!speculative)
spte |= shadow_accessed_mask;
- if (!dirty)
- pte_access &= ~ACC_WRITE_MASK;
+
if (pte_access & ACC_EXEC_MASK)
spte |= shadow_x_mask;
else
@@ -2023,7 +2022,7 @@
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 user_fault, int write_fault,
int *ptwrite, int level, gfn_t gfn,
pfn_t pfn, bool speculative,
bool host_writable)
@@ -2059,7 +2058,7 @@
}
if (set_spte(vcpu, sptep, pte_access, user_fault, write_fault,
- dirty, level, gfn, pfn, speculative, true,
+ level, gfn, pfn, speculative, true,
host_writable)) {
if (write_fault)
*ptwrite = 1;
@@ -2129,7 +2128,7 @@
for (i = 0; i < ret; i++, gfn++, start++)
mmu_set_spte(vcpu, start, ACC_ALL,
- access, 0, 0, 1, NULL,
+ access, 0, 0, NULL,
sp->role.level, gfn,
page_to_pfn(pages[i]), true, true);
@@ -2193,7 +2192,7 @@
unsigned pte_access = ACC_ALL;
mmu_set_spte(vcpu, iterator.sptep, ACC_ALL, pte_access,
- 0, write, 1, &pt_write,
+ 0, write, &pt_write,
level, gfn, pfn, prefault, map_writable);
direct_pte_prefetch(vcpu, iterator.sptep);
++vcpu->stat.pf_fixed;
diff --git a/arch/x86/kvm/paging_tmpl.h b/arch/x86/kvm/paging_tmpl.h
index f0fb1a4..c9fe97b 100644
--- a/arch/x86/kvm/paging_tmpl.h
+++ b/arch/x86/kvm/paging_tmpl.h
@@ -101,11 +101,15 @@
return (ret != orig_pte);
}
-static unsigned FNAME(gpte_access)(struct kvm_vcpu *vcpu, pt_element_t gpte)
+static unsigned FNAME(gpte_access)(struct kvm_vcpu *vcpu, pt_element_t gpte,
+ bool last)
{
unsigned access;
access = (gpte & (PT_WRITABLE_MASK | PT_USER_MASK)) | ACC_EXEC_MASK;
+ if (last && !is_dirty_gpte(gpte))
+ access &= ~ACC_WRITE_MASK;
+
#if PTTYPE == 64
if (vcpu->arch.mmu.nx)
access &= ~(gpte >> PT64_NX_SHIFT);
@@ -232,8 +236,6 @@
pte |= PT_ACCESSED_MASK;
}
- pte_access = pt_access & FNAME(gpte_access)(vcpu, pte);
-
walker->ptes[walker->level - 1] = pte;
if (FNAME(is_last_gpte)(walker, vcpu, mmu, pte)) {
@@ -268,7 +270,7 @@
break;
}
- pt_access = pte_access;
+ pt_access &= FNAME(gpte_access)(vcpu, pte, false);
--walker->level;
}
@@ -293,6 +295,7 @@
walker->ptes[walker->level - 1] = pte;
}
+ pte_access = pt_access & FNAME(gpte_access)(vcpu, pte, true);
walker->pt_access = pt_access;
walker->pte_access = pte_access;
pgprintk("%s: pte %llx pte_access %x pt_access %x\n",
@@ -367,7 +370,7 @@
return;
pgprintk("%s: gpte %llx spte %p\n", __func__, (u64)gpte, spte);
- pte_access = sp->role.access & FNAME(gpte_access)(vcpu, gpte);
+ pte_access = sp->role.access & FNAME(gpte_access)(vcpu, gpte, true);
pfn = gfn_to_pfn_atomic(vcpu->kvm, gpte_to_gfn(gpte));
if (is_error_pfn(pfn)) {
kvm_release_pfn_clean(pfn);
@@ -379,7 +382,7 @@
* vcpu->arch.update_pte.pfn was fetched from get_user_pages(write = 1).
*/
mmu_set_spte(vcpu, spte, sp->role.access, pte_access, 0, 0,
- is_dirty_gpte(gpte), NULL, PT_PAGE_TABLE_LEVEL,
+ NULL, PT_PAGE_TABLE_LEVEL,
gpte_to_gfn(gpte), pfn, true, true);
}
@@ -430,7 +433,6 @@
unsigned pte_access;
gfn_t gfn;
pfn_t pfn;
- bool dirty;
if (spte == sptep)
continue;
@@ -443,18 +445,18 @@
if (FNAME(prefetch_invalid_gpte)(vcpu, sp, spte, gpte))
continue;
- pte_access = sp->role.access & FNAME(gpte_access)(vcpu, gpte);
+ pte_access = sp->role.access & FNAME(gpte_access)(vcpu, gpte,
+ true);
gfn = gpte_to_gfn(gpte);
- dirty = is_dirty_gpte(gpte);
pfn = pte_prefetch_gfn_to_pfn(vcpu, gfn,
- (pte_access & ACC_WRITE_MASK) && dirty);
+ pte_access & ACC_WRITE_MASK);
if (is_error_pfn(pfn)) {
kvm_release_pfn_clean(pfn);
break;
}
mmu_set_spte(vcpu, spte, sp->role.access, pte_access, 0, 0,
- dirty, NULL, PT_PAGE_TABLE_LEVEL, gfn,
+ NULL, PT_PAGE_TABLE_LEVEL, gfn,
pfn, true, true);
}
}
@@ -470,7 +472,6 @@
{
unsigned access = gw->pt_access;
struct kvm_mmu_page *sp = NULL;
- bool dirty = is_dirty_gpte(gw->ptes[gw->level - 1]);
int top_level;
unsigned direct_access;
struct kvm_shadow_walk_iterator it;
@@ -479,8 +480,6 @@
return NULL;
direct_access = gw->pt_access & gw->pte_access;
- if (!dirty)
- direct_access &= ~ACC_WRITE_MASK;
top_level = vcpu->arch.mmu.root_level;
if (top_level == PT32E_ROOT_LEVEL)
@@ -539,7 +538,7 @@
}
mmu_set_spte(vcpu, it.sptep, access, gw->pte_access & access,
- user_fault, write_fault, dirty, ptwrite, it.level,
+ user_fault, write_fault, ptwrite, it.level,
gw->gfn, pfn, prefault, map_writable);
FNAME(pte_prefetch)(vcpu, gw, it.sptep);
@@ -622,17 +621,9 @@
return 0;
/* mmio */
- if (is_error_pfn(pfn)) {
- unsigned access = walker.pte_access;
- bool dirty = is_dirty_gpte(walker.ptes[walker.level - 1]);
-
- if (!dirty)
- access &= ~ACC_WRITE_MASK;
-
+ if (is_error_pfn(pfn))
return kvm_handle_bad_page(vcpu, mmu_is_nested(vcpu) ? 0 :
- addr, access, walker.gfn, pfn);
- }
-
+ addr, walker.pte_access, walker.gfn, pfn);
spin_lock(&vcpu->kvm->mmu_lock);
if (mmu_notifier_retry(vcpu, mmu_seq))
goto out_unlock;
@@ -849,11 +840,12 @@
}
nr_present++;
- pte_access = sp->role.access & FNAME(gpte_access)(vcpu, gpte);
+ pte_access = sp->role.access & FNAME(gpte_access)(vcpu, gpte,
+ true);
host_writable = sp->spt[i] & SPTE_HOST_WRITEABLE;
set_spte(vcpu, &sp->spt[i], pte_access, 0, 0,
- is_dirty_gpte(gpte), PT_PAGE_TABLE_LEVEL, gfn,
+ PT_PAGE_TABLE_LEVEL, gfn,
spte_to_pfn(sp->spt[i]), true, false,
host_writable);
}