Skip to contentSkip to conversation

x86/kvm/mmu: make mmu->prev_roots cache work for NPT case

PATCH Superseded Vitaly Kuznetsov 3 participants 10 messages Feb 22, 2019 – Mar 7, 2019

Subsystems (3)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 x86/kvm/mmu: make mmu->prev_roots cache work for NPT case

I noticed that fast_cr3_switch() always fails when we switch back from L2 to L1 as it is not able to find a cached root. This is odd: host's CR3 usually stays the same, we expect to always follow the fast path. Turns out the problem is that page role is always mismatched because kvm_mmu_get_page() filters out cr4_pae when direct, the value is stored in page header and later compared with new_role in cached_root_available(). As cr4_pae is always set in long mode prev_roots cache is dysfunctional.

The problem appeared after we introduced kvm_calc_mmu_role_common():
previously, we were only setting this bit for shadow MMU root but now we set it for everything. Restore the original behavior.

Fixes 7dcd57552008 ("x86/kvm/mmu: check if tdp/shadow MMU reconfiguration is needed")

---
RFC:
Alternatively, I can suggest two solutions:
- Do not clear cr4_pae in kvm_mmu_get_page() and check direct on call sites (detect_write_misaligned(), get_written_sptes()).
- Filter cr4_pae out when direct in kvm_mmu_new_cr3().
The code in kvm_mmu_get_page() is very ancient, I'm afraid to touch it :=)

---
 arch/x86/kvm/mmu.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
arch/x86/kvm/mmu.c +2 -1
index f2d1d230d5b8..c729e98eee49 100644
@@ -4791,7 +4791,6 @@ static union kvm_mmu_role kvm_calc_mmu_role_common(struct kvm_vcpu *vcpu,
 
 	role.base.access = ACC_ALL;
 	role.base.nxe = !!is_nx(vcpu);
-	role.base.cr4_pae = !!is_pae(vcpu);
 	role.base.cr0_wp = is_write_protection(vcpu);
 	role.base.smm = is_smm(vcpu);
 	role.base.guest_mode = is_guest_mode(vcpu);
@@ -4871,6 +4870,8 @@ kvm_calc_shadow_mmu_root_page_role(struct kvm_vcpu *vcpu, bool base_only)
 {
 	union kvm_mmu_role role = kvm_calc_mmu_role_common(vcpu, base_only);
 
+	role.base.cr4_pae = !!is_pae(vcpu);
+
 	role.base.smep_andnot_wp = role.ext.cr4_smep &&
 		!is_write_protection(vcpu);
 	role.base.smap_andnot_wp = role.ext.cr4_smap &&
Signature · 1 line
2.20.1
maintainer
#1 lore raw
Quoted 7 lines · Vitaly Kuznetsov

I noticed that fast_cr3_switch() always fails when we switch back from L2 to L1 as it is not able to find a cached root. This is odd: host's CR3 usually stays the same, we expect to always follow the fast path. Turns out the problem is that page role is always mismatched because kvm_mmu_get_page() filters out cr4_pae when direct, the value is stored in page header and later compared with new_role in cached_root_available(). As cr4_pae is always set in long mode prev_roots cache is dysfunctional.

Really cr4_pae means "are the PTEs 8 bytes". So I think your patch is correct but on top we should set it to 1 (not zero!!) for kvm_calc_shadow_ept_root_page_role, init_kvm_nested_mmu and kvm_calc_tdp_mmu_root_page_role. Or maybe everything breaks with that change.

Quoted 2 lines

- Do not clear cr4_pae in kvm_mmu_get_page() and check direct on call sites (detect_write_misaligned(), get_written_sptes()).

These only run with shadow page tables, by the way.

Paolo

#2 lore raw
Quoted 9 lines · Paolo Bonzini
Quoted 7 lines · Vitaly Kuznetsov

I noticed that fast_cr3_switch() always fails when we switch back from L2 to L1 as it is not able to find a cached root. This is odd: host's CR3 usually stays the same, we expect to always follow the fast path. Turns out the problem is that page role is always mismatched because kvm_mmu_get_page() filters out cr4_pae when direct, the value is stored in page header and later compared with new_role in cached_root_available(). As cr4_pae is always set in long mode prev_roots cache is dysfunctional.

Quoted 6 lines

Really cr4_pae means "are the PTEs 8 bytes". So I think your patch is correct but on top we should set it to 1 (not zero!!) for kvm_calc_shadow_ept_root_page_role, init_kvm_nested_mmu and kvm_calc_tdp_mmu_root_page_role. Or maybe everything breaks with that change.

Yes, exactly. If we put '1' there kvm_mmu_get_page() will again filter it out and we won't be able to find the root in prev_roots cache :-(

Quoted 5 lines
Quoted 2 lines

- Do not clear cr4_pae in kvm_mmu_get_page() and check direct on call sites (detect_write_misaligned(), get_written_sptes()).

These only run with shadow page tables, by the way.

Yes, and that's why I think it may make sense to move the filtering logic there. At least in other cases cr4_pae will always be equal to is_pae().

It seems I know too little about shadow paging and all these corner cases :-(

Signature · 1 line
Vitaly
maintainer
#3 lore raw
Quoted 9 lines · Vitaly Kuznetsov
Quoted 6 lines

Really cr4_pae means "are the PTEs 8 bytes". So I think your patch is correct but on top we should set it to 1 (not zero!!) for kvm_calc_shadow_ept_root_page_role, init_kvm_nested_mmu and kvm_calc_tdp_mmu_root_page_role. Or maybe everything breaks with that change.

Yes, exactly. If we put '1' there kvm_mmu_get_page() will again filter it out and we won't be able to find the root in prev_roots cache :-(

Well, of course then kvm_mmu_get_page() would have to remove the filtering.

Paolo

#4 lore raw

RFC PATCH v2 1/1 x86/kvm/mmu: make mmu->prev_roots cache work for NPT case

Alternative patch: remove the filtering from kvm_mmu_get_page() and check for direct on call sites. cr4_pae setting in kvm_calc_mmu_role_common() can be preserved for consistency.

---
 arch/x86/kvm/mmu.c | 6 ++----
 1 file changed, 2 insertions(+), 4 deletions(-)
arch/x86/kvm/mmu.c +2 -4
index f2d1d230d5b8..7fb8118f2af6 100644
@@ -2420,8 +2420,6 @@ static struct kvm_mmu_page *kvm_mmu_get_page(struct kvm_vcpu *vcpu,
 	role = vcpu->arch.mmu->mmu_role.base;
 	role.level = level;
 	role.direct = direct;
-	if (role.direct)
-		role.cr4_pae = 0;
 	role.access = access;
 	if (!vcpu->arch.mmu->direct_map
 	    && vcpu->arch.mmu->root_level <= PT32_ROOT_LEVEL) {
@@ -5176,7 +5174,7 @@ static bool detect_write_misaligned(struct kvm_mmu_page *sp, gpa_t gpa,
 		 gpa, bytes, sp->role.word);
 
 	offset = offset_in_page(gpa);
-	pte_size = sp->role.cr4_pae ? 8 : 4;
+	pte_size = (sp->role.cr4_pae && !sp->role.direct) ? 8 : 4;
 
 	/*
 	 * Sometimes, the OS only writes the last one bytes to update status
@@ -5200,7 +5198,7 @@ static u64 *get_written_sptes(struct kvm_mmu_page *sp, gpa_t gpa, int *nspte)
 	page_offset = offset_in_page(gpa);
 	level = sp->role.level;
 	*nspte = 1;
-	if (!sp->role.cr4_pae) {
+	if (!sp->role.cr4_pae || sp->role.direct) {
 		page_offset <<= 1;	/* 32->64 */
 		/*
 		 * A 32-bit pde maps 4MB while the shadow pdes map
Signature · 1 line
2.20.1
#5 lore raw
Quoted 32 lines · Vitaly Kuznetsov

Alternative patch: remove the filtering from kvm_mmu_get_page() and check for direct on call sites. cr4_pae setting in kvm_calc_mmu_role_common() can be preserved for consistency.

---
 arch/x86/kvm/mmu.c | 6 ++----
 1 file changed, 2 insertions(+), 4 deletions(-)
arch/x86/kvm/mmu.c +1 -3
index f2d1d230d5b8..7fb8118f2af6 100644
@@ -2420,8 +2420,6 @@ static struct kvm_mmu_page *kvm_mmu_get_page(struct kvm_vcpu *vcpu,
 	role = vcpu->arch.mmu->mmu_role.base;
 	role.level = level;
 	role.direct = direct;
-	if (role.direct)
-		role.cr4_pae = 0;
 	role.access = access;
 	if (!vcpu->arch.mmu->direct_map
 	    && vcpu->arch.mmu->root_level <= PT32_ROOT_LEVEL) {
@@ -5176,7 +5174,7 @@ static bool detect_write_misaligned(struct kvm_mmu_page *sp, gpa_t gpa,
 		 gpa, bytes, sp->role.word);
 
 	offset = offset_in_page(gpa);
-	pte_size = sp->role.cr4_pae ? 8 : 4;
+	pte_size = (sp->role.cr4_pae && !sp->role.direct) ? 8 : 4;
 
 	/*
 	 * Sometimes, the OS only writes the last one bytes to update status

arch/x86/kvm/mmu.c static u64 *get_written_sptes(struct kvm_mmu_page *sp, gpa_t gpa, int *nspte) · line 5204

Quoted 8 lines
arch/x86/kvm/mmu.c +1 -1
@@ -5200,7 +5198,7 @@ static u64 *get_written_sptes(struct kvm_mmu_page *sp, gpa_t gpa, int *nspte)
 	page_offset = offset_in_page(gpa);
 	level = sp->role.level;
 	*nspte = 1;
-	if (!sp->role.cr4_pae) {
+	if (!sp->role.cr4_pae || sp->role.direct) {
 		page_offset <<= 1;	/* 32->64 */
 		/*
 		 * A 32-bit pde maps 4MB while the shadow pdes map

While I personally prefer this approach to not setting role.cr4_pae in kvm_calc_mmu_role_common() I'd like to get maintainers opinion (I did test the patch with ept=off but I have to admit I don't know much about shadow page tables which actually use detect_write_misaligned()/ get_written_sptes())

Paolo, Radim, (anyone else) - any thoughts?

Signature · 1 line
Vitaly
#6 lore raw
Quoted 49 lines · Vitaly Kuznetsov
Quoted 32 lines · Vitaly Kuznetsov

Alternative patch: remove the filtering from kvm_mmu_get_page() and check for direct on call sites. cr4_pae setting in kvm_calc_mmu_role_common() can be preserved for consistency.

---
 arch/x86/kvm/mmu.c | 6 ++----
 1 file changed, 2 insertions(+), 4 deletions(-)
arch/x86/kvm/mmu.c +1 -3
index f2d1d230d5b8..7fb8118f2af6 100644
@@ -2420,8 +2420,6 @@ static struct kvm_mmu_page *kvm_mmu_get_page(struct kvm_vcpu *vcpu,
 	role = vcpu->arch.mmu->mmu_role.base;
 	role.level = level;
 	role.direct = direct;
-	if (role.direct)
-		role.cr4_pae = 0;
 	role.access = access;
 	if (!vcpu->arch.mmu->direct_map
 	    && vcpu->arch.mmu->root_level <= PT32_ROOT_LEVEL) {
@@ -5176,7 +5174,7 @@ static bool detect_write_misaligned(struct kvm_mmu_page *sp, gpa_t gpa,
 		 gpa, bytes, sp->role.word);
 
 	offset = offset_in_page(gpa);
-	pte_size = sp->role.cr4_pae ? 8 : 4;
+	pte_size = (sp->role.cr4_pae && !sp->role.direct) ? 8 : 4;
 
 	/*
 	 * Sometimes, the OS only writes the last one bytes to update status

arch/x86/kvm/mmu.c static u64 *get_written_sptes(struct kvm_mmu_page *sp, gpa_t gpa, int *nspte) · line 5204

Quoted 8 lines
arch/x86/kvm/mmu.c +1 -1
@@ -5200,7 +5198,7 @@ static u64 *get_written_sptes(struct kvm_mmu_page *sp, gpa_t gpa, int *nspte)
 	page_offset = offset_in_page(gpa);
 	level = sp->role.level;
 	*nspte = 1;
-	if (!sp->role.cr4_pae) {
+	if (!sp->role.cr4_pae || sp->role.direct) {
 		page_offset <<= 1;	/* 32->64 */
 		/*
 		 * A 32-bit pde maps 4MB while the shadow pdes map

While I personally prefer this approach to not setting role.cr4_pae in kvm_calc_mmu_role_common() I'd like to get maintainers opinion (I did test the patch with ept=off but I have to admit I don't know much about shadow page tables which actually use detect_write_misaligned()/ get_written_sptes())

Quoted 1 line

Paolo, Radim, (anyone else) - any thoughts?

The changes to detect_write_misaligned() and get_written_sptes() are wrong/unnecessary as those functions should only be called for indirect shadow PTEs, e.g.:

	for_each_gfn_indirect_valid_sp(vcpu->kvm, sp, gfn) {
		if (detect_write_misaligned(sp, gpa, bytes) ||
		      detect_write_flooding(sp)) {
			kvm_mmu_prepare_zap_page(vcpu->kvm, sp, &invalid_list);
			++vcpu->kvm->stat.mmu_flooded;
			continue;
		}

		spte = get_written_sptes(sp, gpa, &npte);
		if (!spte)
			continue;

		...
	}

If anything, they could WARN_ON(sp->role.direct), but IMO that's overkill since they're static helpers with a single call site (above).

I'm missing the background for this patch, why does clearing role.cr4_pae for direct SPTEs cause problems with the prev_roots cache?

#7 lore raw
Quoted 71 lines · Sean Christopherson
Quoted 49 lines · Vitaly Kuznetsov
Quoted 32 lines · Vitaly Kuznetsov

Alternative patch: remove the filtering from kvm_mmu_get_page() and check for direct on call sites. cr4_pae setting in kvm_calc_mmu_role_common() can be preserved for consistency.

---
 arch/x86/kvm/mmu.c | 6 ++----
 1 file changed, 2 insertions(+), 4 deletions(-)
arch/x86/kvm/mmu.c +1 -3
index f2d1d230d5b8..7fb8118f2af6 100644
@@ -2420,8 +2420,6 @@ static struct kvm_mmu_page *kvm_mmu_get_page(struct kvm_vcpu *vcpu,
 	role = vcpu->arch.mmu->mmu_role.base;
 	role.level = level;
 	role.direct = direct;
-	if (role.direct)
-		role.cr4_pae = 0;
 	role.access = access;
 	if (!vcpu->arch.mmu->direct_map
 	    && vcpu->arch.mmu->root_level <= PT32_ROOT_LEVEL) {
@@ -5176,7 +5174,7 @@ static bool detect_write_misaligned(struct kvm_mmu_page *sp, gpa_t gpa,
 		 gpa, bytes, sp->role.word);
 
 	offset = offset_in_page(gpa);
-	pte_size = sp->role.cr4_pae ? 8 : 4;
+	pte_size = (sp->role.cr4_pae && !sp->role.direct) ? 8 : 4;
 
 	/*
 	 * Sometimes, the OS only writes the last one bytes to update status

arch/x86/kvm/mmu.c static u64 *get_written_sptes(struct kvm_mmu_page *sp, gpa_t gpa, int *nspte) · line 5204

Quoted 8 lines
arch/x86/kvm/mmu.c +1 -1
@@ -5200,7 +5198,7 @@ static u64 *get_written_sptes(struct kvm_mmu_page *sp, gpa_t gpa, int *nspte)
 	page_offset = offset_in_page(gpa);
 	level = sp->role.level;
 	*nspte = 1;
-	if (!sp->role.cr4_pae) {
+	if (!sp->role.cr4_pae || sp->role.direct) {
 		page_offset <<= 1;	/* 32->64 */
 		/*
 		 * A 32-bit pde maps 4MB while the shadow pdes map

While I personally prefer this approach to not setting role.cr4_pae in kvm_calc_mmu_role_common() I'd like to get maintainers opinion (I did test the patch with ept=off but I have to admit I don't know much about shadow page tables which actually use detect_write_misaligned()/ get_written_sptes())

Quoted 1 line

Paolo, Radim, (anyone else) - any thoughts?

The changes to detect_write_misaligned() and get_written_sptes() are wrong/unnecessary as those functions should only be called for indirect shadow PTEs, e.g.:

	for_each_gfn_indirect_valid_sp(vcpu->kvm, sp, gfn) {
		if (detect_write_misaligned(sp, gpa, bytes) ||
		      detect_write_flooding(sp)) {
			kvm_mmu_prepare_zap_page(vcpu->kvm, sp, &invalid_list);
			++vcpu->kvm->stat.mmu_flooded;
			continue;
		}

		spte = get_written_sptes(sp, gpa, &npte);
		if (!spte)
			continue;

		...
	}
Quoted 5 lines

If anything, they could WARN_ON(sp->role.direct), but IMO that's overkill since they're static helpers with a single call site (above).

I'm missing the background for this patch, why does clearing role.cr4_pae for direct SPTEs cause problems with the prev_roots cache?

Oh, thank you for taking a look! You've probably missed my original v1 patch where I tried to explaing the issue:

"I noticed that fast_cr3_switch() always fails when we switch back from L2 to L1 as it is not able to find a cached root. This is odd: host's CR3 usually stays the same, we expect to always follow the fast path. Turns out the problem is that page role is always mismatched because kvm_mmu_get_page() filters out cr4_pae when direct, the value is stored in page header and later compared with new_role in cached_root_available(). As cr4_pae is always set in long mode prev_roots cache is dysfunctional.

The problem appeared after we introduced kvm_calc_mmu_role_common():
previously, we were only setting this bit for shadow MMU root but now we set it for everything. Restore the original behavior.

Fixes 7dcd57552008 ("x86/kvm/mmu: check if tdp/shadow MMU reconfiguration is needed")

"

and the patch was just removing role.cr4_pae setting and moving it to kvm_calc_shadow_mmu_root_page_role(). This is an alternative approach - always set cr4_pae but count on kvm_mmu_get_page() filtering out cr4_pae when direct is set.

Signature · 1 line
Vitaly
#8 lore raw
Quoted 95 lines · Vitaly Kuznetsov
Quoted 71 lines · Sean Christopherson
Quoted 49 lines · Vitaly Kuznetsov
Quoted 32 lines · Vitaly Kuznetsov

Alternative patch: remove the filtering from kvm_mmu_get_page() and check for direct on call sites. cr4_pae setting in kvm_calc_mmu_role_common() can be preserved for consistency.

---
 arch/x86/kvm/mmu.c | 6 ++----
 1 file changed, 2 insertions(+), 4 deletions(-)
arch/x86/kvm/mmu.c +1 -3
index f2d1d230d5b8..7fb8118f2af6 100644
@@ -2420,8 +2420,6 @@ static struct kvm_mmu_page *kvm_mmu_get_page(struct kvm_vcpu *vcpu,
 	role = vcpu->arch.mmu->mmu_role.base;
 	role.level = level;
 	role.direct = direct;
-	if (role.direct)
-		role.cr4_pae = 0;
 	role.access = access;
 	if (!vcpu->arch.mmu->direct_map
 	    && vcpu->arch.mmu->root_level <= PT32_ROOT_LEVEL) {
@@ -5176,7 +5174,7 @@ static bool detect_write_misaligned(struct kvm_mmu_page *sp, gpa_t gpa,
 		 gpa, bytes, sp->role.word);
 
 	offset = offset_in_page(gpa);
-	pte_size = sp->role.cr4_pae ? 8 : 4;
+	pte_size = (sp->role.cr4_pae && !sp->role.direct) ? 8 : 4;
 
 	/*
 	 * Sometimes, the OS only writes the last one bytes to update status

arch/x86/kvm/mmu.c static u64 *get_written_sptes(struct kvm_mmu_page *sp, gpa_t gpa, int *nspte) · line 5204

Quoted 8 lines
arch/x86/kvm/mmu.c +1 -1
@@ -5200,7 +5198,7 @@ static u64 *get_written_sptes(struct kvm_mmu_page *sp, gpa_t gpa, int *nspte)
 	page_offset = offset_in_page(gpa);
 	level = sp->role.level;
 	*nspte = 1;
-	if (!sp->role.cr4_pae) {
+	if (!sp->role.cr4_pae || sp->role.direct) {
 		page_offset <<= 1;	/* 32->64 */
 		/*
 		 * A 32-bit pde maps 4MB while the shadow pdes map

While I personally prefer this approach to not setting role.cr4_pae in kvm_calc_mmu_role_common() I'd like to get maintainers opinion (I did test the patch with ept=off but I have to admit I don't know much about shadow page tables which actually use detect_write_misaligned()/ get_written_sptes())

Quoted 1 line

Paolo, Radim, (anyone else) - any thoughts?

The changes to detect_write_misaligned() and get_written_sptes() are wrong/unnecessary as those functions should only be called for indirect shadow PTEs, e.g.:

	for_each_gfn_indirect_valid_sp(vcpu->kvm, sp, gfn) {
		if (detect_write_misaligned(sp, gpa, bytes) ||
		      detect_write_flooding(sp)) {
			kvm_mmu_prepare_zap_page(vcpu->kvm, sp, &invalid_list);
			++vcpu->kvm->stat.mmu_flooded;
			continue;
		}

		spte = get_written_sptes(sp, gpa, &npte);
		if (!spte)
			continue;

		...
	}
Quoted 5 lines

If anything, they could WARN_ON(sp->role.direct), but IMO that's overkill since they're static helpers with a single call site (above).

I'm missing the background for this patch, why does clearing role.cr4_pae for direct SPTEs cause problems with the prev_roots cache?

Oh, thank you for taking a look! You've probably missed my original v1 patch where I tried to explaing the issue:

"I noticed that fast_cr3_switch() always fails when we switch back from L2 to L1 as it is not able to find a cached root. This is odd: host's CR3 usually stays the same, we expect to always follow the fast path. Turns out the problem is that page role is always mismatched because kvm_mmu_get_page() filters out cr4_pae when direct, the value is stored in page header and later compared with new_role in cached_root_available(). As cr4_pae is always set in long mode prev_roots cache is dysfunctional.

The problem appeared after we introduced kvm_calc_mmu_role_common():
previously, we were only setting this bit for shadow MMU root but now we set it for everything. Restore the original behavior.

Quoted 7 lines
Fixes 7dcd57552008 ("x86/kvm/mmu: check if tdp/shadow MMU reconfiguration is needed")

"

and the patch was just removing role.cr4_pae setting and moving it to kvm_calc_shadow_mmu_root_page_role(). This is an alternative approach - always set cr4_pae but count on kvm_mmu_get_page() filtering out cr4_pae when direct is set.

Ah, but neither approach solves the underlying problem that KVM can't identify a shadow page that is associated with a nested EPT entry. For example, kvm_calc_shadow_ept_root_page_role() should set cr4_pae=1, i.e. always has 8-byte guest PTEs regardless of PAE. But that would break e.g. __kvm_sync_page() as it would always mismatch if L2 is a non-PAE guest (I think I got that right...).

I think what we could do is repurpose role's nxe, cr0_wp, and sm{a,e}p_andnot_wp bits to uniquely identify a nested EPT/NPT entry. E.g. cr0_wp=1 and sm{a,e}p_andnot_wp=1 are an impossible combination. I'll throw together a patch to see what breaks. In fact, I think we could revamp kvm_calc_shadow_ept_root_page_role() to completely ignore all legacy paging bits, i.e. handling changes in L2's configuration is L1's problem.

#9 lore raw
Quoted 2 lines · Sean Christopherson

I think what we could do is repurpose role's nxe, cr0_wp, and sm{a,e}p_andnot_wp bits to uniquely identify a nested EPT/NPT entry.

Ignore the "NPT" comment, this would only apply to EPT.

Quoted 5 lines

E.g. cr0_wp=1 and sm{a,e}p_andnot_wp=1 are an impossible combination. I'll throw together a patch to see what breaks. In fact, I think we could revamp kvm_calc_shadow_ept_root_page_role() to completely ignore all legacy paging bits, i.e. handling changes in L2's configuration is L1's problem.