Skip to contentSkip to conversation

KVM: MMU: Fix a refcount bug with ZONE_DEVICE pages

PATCH Applied Sean Christopherson 4 participants 8 messages Nov 11, 2019 – Nov 12, 2019

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

v2 cover KVM: MMU: Fix a refcount bug with ZONE_DEVICE pages

This mini-series fixes a suspected, but technically unconfirmed, bug in KVM related to ZONE_DEVICE pages. The suspected issue is that KVM treats ZONE_DEVICE pages as reserved PFNs, and so doesn't put references to such pages when dropping references via KVM's generic kvm_release_pfn_clean().

David Hildenbrand uncovered the bug during a discussion about removing PG_reserved from ZONE_DEVICE pages, after Dan Williams pointed out[1] that there was a bug report from Adam Borowski[2] that was likely related to KVM's interaction with PageReserved().

Patch 1/3 contains the actual fix, patches 2/3 and 3/3 are minor cleanup that is mostly unrelated, but dependent and prompted by the fix itself.

v2:
- Remove the kvm_is_zone_device_pfn(pfn) check from kvm_get_pfn(). It's not entirely clear whether or not the hva_to_pfn_remapped() case is actually broken, e.g. KVM's page fault handler is likely ok, whereas not calling get_page() willl definitely cause breakage as KVM would later call put_page() on the pfn/page. [Paolo]

- WARN if kvm_is_zone_device_pfn() is called without the underlying page being pinned. This won't necessarily catch all bugs, e.g. if the above hva_to_pfn_remapped case is indeed broken, but will prevent completely bogus usage. [Dan]

- Remove the is_error_pfn() check from transparent_hugepage_adjust() instead of carrying it forward into the new kvm_is_hugepage_allowed() helper. [Paolo]

[1] http://lkml.kernel.org/r/[email protected]
[2] https://lkml.kernel.org/r/[email protected]

Sean Christopherson (3):
KVM: MMU: Do not treat ZONE_DEVICE pages as being reserved
KVM: x86/mmu: Remove superfluous is_error_pfn() check from THP adjust
KVM: x86/mmu: Add helper to consolidate huge page promotion

 arch/x86/kvm/mmu.c       | 15 +++++++++------
 include/linux/kvm_host.h |  1 +
 virt/kvm/kvm_main.c      | 26 +++++++++++++++++++++++---
 3 files changed, 33 insertions(+), 9 deletions(-)
Signature · 1 line
2.24.0
#1 lore raw

PATCH v2 1/3 KVM: MMU: Do not treat ZONE_DEVICE pages as being reserved

Explicitly exempt ZONE_DEVICE pages from kvm_is_reserved_pfn() and instead manually handle ZONE_DEVICE on a case-by-case basis. For things like page refcounts, KVM needs to treat ZONE_DEVICE pages like normal pages, e.g. put pages grabbed via gup(). But for flows such as setting A/D bits or shifting refcounts for transparent huge pages, KVM needs to to avoid processing ZONE_DEVICE pages as the flows in question lack the underlying machinery for proper handling of ZONE_DEVICE pages.

This fixes a hang reported by Adam Borowski[*] in dev_pagemap_cleanup() when running a KVM guest backed with /dev/dax memory, as KVM straight up doesn't put any references to ZONE_DEVICE pages acquired by gup().

Note, Dan Williams proposed an alternative solution of doing put_page() on ZONE_DEVICE pages immediately after gup() in order to simplify the auditing needed to ensure is_zone_device_page() is called if and only if the backing device is pinned (via gup()). But that approach would break kvm_vcpu_{un}map() as KVM requires the page to be pinned from map() 'til unmap() when accessing guest memory, unlike KVM's secondary MMU, which coordinates with mmu_notifier invalidations to avoid creating stale page references, i.e. doesn't rely on pages being pinned.

[*] http://lkml.kernel.org/r/[email protected]

---
 arch/x86/kvm/mmu.c       |  8 ++++----
 include/linux/kvm_host.h |  1 +
 virt/kvm/kvm_main.c      | 26 +++++++++++++++++++++++---
 3 files changed, 28 insertions(+), 7 deletions(-)
arch/x86/kvm/mmu.c +4 -4
index 24c23c66b226..bf82b1f2e834 100644
@@ -3306,7 +3306,7 @@ static void transparent_hugepage_adjust(struct kvm_vcpu *vcpu,
 	 * here.
 	 */
 	if (!is_error_noslot_pfn(pfn) && !kvm_is_reserved_pfn(pfn) &&
-	    level == PT_PAGE_TABLE_LEVEL &&
+	    !kvm_is_zone_device_pfn(pfn) && level == PT_PAGE_TABLE_LEVEL &&
 	    PageTransCompoundMap(pfn_to_page(pfn)) &&
 	    !mmu_gfn_lpage_is_disallowed(vcpu, gfn, PT_DIRECTORY_LEVEL)) {
 		unsigned long mask;
@@ -5914,9 +5914,9 @@ static bool kvm_mmu_zap_collapsible_spte(struct kvm *kvm,
 		 * the guest, and the guest page table is using 4K page size
 		 * mapping if the indirect sp has level = 1.
 		 */
-		if (sp->role.direct &&
-			!kvm_is_reserved_pfn(pfn) &&
-			PageTransCompoundMap(pfn_to_page(pfn))) {
+		if (sp->role.direct && !kvm_is_reserved_pfn(pfn) &&
+		    !kvm_is_zone_device_pfn(pfn) &&
+		    PageTransCompoundMap(pfn_to_page(pfn))) {
 			pte_list_remove(rmap_head, sptep);
 
 			if (kvm_available_flush_tlb_with_range())
include/linux/kvm_host.h +1 -0
index a817e446c9aa..4ad1cd7d2d4d 100644
@@ -966,6 +966,7 @@ int kvm_cpu_has_pending_timer(struct kvm_vcpu *vcpu);
 void kvm_vcpu_kick(struct kvm_vcpu *vcpu);
 
 bool kvm_is_reserved_pfn(kvm_pfn_t pfn);
+bool kvm_is_zone_device_pfn(kvm_pfn_t pfn);
 
 struct kvm_irq_ack_notifier {
 	struct hlist_node link;
virt/kvm/kvm_main.c +23 -3
index b8534c6b8cf6..bc9d10a0a334 100644
@@ -149,10 +149,30 @@ __weak int kvm_arch_mmu_notifier_invalidate_range(struct kvm *kvm,
 	return 0;
 }
 
+bool kvm_is_zone_device_pfn(kvm_pfn_t pfn)
+{
+	/*
+	 * The metadata used by is_zone_device_page() to determine whether or
+	 * not a page is ZONE_DEVICE is guaranteed to be valid if and only if
+	 * the device has been pinned, e.g. by get_user_pages().  WARN if the
+	 * page_count() is zero to help detect bad usage of this helper.
+	 */
+	if (!pfn_valid(pfn) || WARN_ON_ONCE(!page_count(pfn_to_page(pfn))))
+		return false;
+
+	return is_zone_device_page(pfn_to_page(pfn));
+}
+
 bool kvm_is_reserved_pfn(kvm_pfn_t pfn)
 {
+	/*
+	 * ZONE_DEVICE pages currently set PG_reserved, but from a refcounting
+	 * perspective they are "normal" pages, albeit with slightly different
+	 * usage rules.
+	 */
 	if (pfn_valid(pfn))
-		return PageReserved(pfn_to_page(pfn));
+		return PageReserved(pfn_to_page(pfn)) &&
+		       !kvm_is_zone_device_pfn(pfn);
 
 	return true;
 }
@@ -1865,7 +1885,7 @@ EXPORT_SYMBOL_GPL(kvm_release_pfn_dirty);
 
 void kvm_set_pfn_dirty(kvm_pfn_t pfn)
 {
-	if (!kvm_is_reserved_pfn(pfn)) {
+	if (!kvm_is_reserved_pfn(pfn) && !kvm_is_zone_device_pfn(pfn)) {
 		struct page *page = pfn_to_page(pfn);
 
 		SetPageDirty(page);
@@ -1875,7 +1895,7 @@ EXPORT_SYMBOL_GPL(kvm_set_pfn_dirty);
 
 void kvm_set_pfn_accessed(kvm_pfn_t pfn)
 {
-	if (!kvm_is_reserved_pfn(pfn))
+	if (!kvm_is_reserved_pfn(pfn) && !kvm_is_zone_device_pfn(pfn))
 		mark_page_accessed(pfn_to_page(pfn));
 }
 EXPORT_SYMBOL_GPL(kvm_set_pfn_accessed);
Signature · 1 line
2.24.0
maintainer Applied
#2 lore raw

Re: [PATCH v2 1/3] KVM: MMU: Do not treat ZONE_DEVICE pages as being reserved

Quoted 119 lines · Sean Christopherson

Explicitly exempt ZONE_DEVICE pages from kvm_is_reserved_pfn() and instead manually handle ZONE_DEVICE on a case-by-case basis. For things like page refcounts, KVM needs to treat ZONE_DEVICE pages like normal pages, e.g. put pages grabbed via gup(). But for flows such as setting A/D bits or shifting refcounts for transparent huge pages, KVM needs to to avoid processing ZONE_DEVICE pages as the flows in question lack the underlying machinery for proper handling of ZONE_DEVICE pages.

This fixes a hang reported by Adam Borowski[*] in dev_pagemap_cleanup() when running a KVM guest backed with /dev/dax memory, as KVM straight up doesn't put any references to ZONE_DEVICE pages acquired by gup().

Note, Dan Williams proposed an alternative solution of doing put_page() on ZONE_DEVICE pages immediately after gup() in order to simplify the auditing needed to ensure is_zone_device_page() is called if and only if the backing device is pinned (via gup()). But that approach would break kvm_vcpu_{un}map() as KVM requires the page to be pinned from map() 'til unmap() when accessing guest memory, unlike KVM's secondary MMU, which coordinates with mmu_notifier invalidations to avoid creating stale page references, i.e. doesn't rely on pages being pinned.

[*] http://lkml.kernel.org/r/[email protected]

---
 arch/x86/kvm/mmu.c       |  8 ++++----
 include/linux/kvm_host.h |  1 +
 virt/kvm/kvm_main.c      | 26 +++++++++++++++++++++++---
 3 files changed, 28 insertions(+), 7 deletions(-)
arch/x86/kvm/mmu.c +4 -4
index 24c23c66b226..bf82b1f2e834 100644
@@ -3306,7 +3306,7 @@ static void transparent_hugepage_adjust(struct kvm_vcpu *vcpu,
 	 * here.
 	 */
 	if (!is_error_noslot_pfn(pfn) && !kvm_is_reserved_pfn(pfn) &&
-	    level == PT_PAGE_TABLE_LEVEL &&
+	    !kvm_is_zone_device_pfn(pfn) && level == PT_PAGE_TABLE_LEVEL &&
 	    PageTransCompoundMap(pfn_to_page(pfn)) &&
 	    !mmu_gfn_lpage_is_disallowed(vcpu, gfn, PT_DIRECTORY_LEVEL)) {
 		unsigned long mask;
@@ -5914,9 +5914,9 @@ static bool kvm_mmu_zap_collapsible_spte(struct kvm *kvm,
 		 * the guest, and the guest page table is using 4K page size
 		 * mapping if the indirect sp has level = 1.
 		 */
-		if (sp->role.direct &&
-			!kvm_is_reserved_pfn(pfn) &&
-			PageTransCompoundMap(pfn_to_page(pfn))) {
+		if (sp->role.direct && !kvm_is_reserved_pfn(pfn) &&
+		    !kvm_is_zone_device_pfn(pfn) &&
+		    PageTransCompoundMap(pfn_to_page(pfn))) {
 			pte_list_remove(rmap_head, sptep);
 
 			if (kvm_available_flush_tlb_with_range())
include/linux/kvm_host.h +1 -0
index a817e446c9aa..4ad1cd7d2d4d 100644
@@ -966,6 +966,7 @@ int kvm_cpu_has_pending_timer(struct kvm_vcpu *vcpu);
 void kvm_vcpu_kick(struct kvm_vcpu *vcpu);
 
 bool kvm_is_reserved_pfn(kvm_pfn_t pfn);
+bool kvm_is_zone_device_pfn(kvm_pfn_t pfn);
 
 struct kvm_irq_ack_notifier {
 	struct hlist_node link;
virt/kvm/kvm_main.c +22 -2
index b8534c6b8cf6..bc9d10a0a334 100644
@@ -149,10 +149,30 @@ __weak int kvm_arch_mmu_notifier_invalidate_range(struct kvm *kvm,
 	return 0;
 }
 
+bool kvm_is_zone_device_pfn(kvm_pfn_t pfn)
+{
+	/*
+	 * The metadata used by is_zone_device_page() to determine whether or
+	 * not a page is ZONE_DEVICE is guaranteed to be valid if and only if
+	 * the device has been pinned, e.g. by get_user_pages().  WARN if the
+	 * page_count() is zero to help detect bad usage of this helper.
+	 */
+	if (!pfn_valid(pfn) || WARN_ON_ONCE(!page_count(pfn_to_page(pfn))))
+		return false;
+
+	return is_zone_device_page(pfn_to_page(pfn));
+}
+
 bool kvm_is_reserved_pfn(kvm_pfn_t pfn)
 {
+	/*
+	 * ZONE_DEVICE pages currently set PG_reserved, but from a refcounting
+	 * perspective they are "normal" pages, albeit with slightly different
+	 * usage rules.
+	 */
 	if (pfn_valid(pfn))
-		return PageReserved(pfn_to_page(pfn));
+		return PageReserved(pfn_to_page(pfn)) &&
+		       !kvm_is_zone_device_pfn(pfn);
 
 	return true;
 }
@@ -1865,7 +1885,7 @@ EXPORT_SYMBOL_GPL(kvm_release_pfn_dirty);
 
 void kvm_set_pfn_dirty(kvm_pfn_t pfn)
 {
-	if (!kvm_is_reserved_pfn(pfn)) {
+	if (!kvm_is_reserved_pfn(pfn) && !kvm_is_zone_device_pfn(pfn)) {
 		struct page *page = pfn_to_page(pfn);
 
 		SetPageDirty(page);

virt/kvm/kvm_main.c EXPORT_SYMBOL_GPL(kvm_set_pfn_dirty); · line 1901

Quoted 8 lines
virt/kvm/kvm_main.c +1 -1
@@ -1875,7 +1895,7 @@ EXPORT_SYMBOL_GPL(kvm_set_pfn_dirty);
 
 void kvm_set_pfn_accessed(kvm_pfn_t pfn)
 {
-	if (!kvm_is_reserved_pfn(pfn))
+	if (!kvm_is_reserved_pfn(pfn) && !kvm_is_zone_device_pfn(pfn))
 		mark_page_accessed(pfn_to_page(pfn));
 }
 EXPORT_SYMBOL_GPL(kvm_set_pfn_accessed);

Queued, thanks -- the other two will wait for after the merge window to avoid pointless conflicts.

Paolo

Acked-by on v2 1/3
#3 lore raw

Re: [PATCH v2 1/3] KVM: MMU: Do not treat ZONE_DEVICE pages as being reserved

In reply to Sean Christopherson
Quoted 22 lines · Sean Christopherson

Explicitly exempt ZONE_DEVICE pages from kvm_is_reserved_pfn() and instead manually handle ZONE_DEVICE on a case-by-case basis. For things like page refcounts, KVM needs to treat ZONE_DEVICE pages like normal pages, e.g. put pages grabbed via gup(). But for flows such as setting A/D bits or shifting refcounts for transparent huge pages, KVM needs to to avoid processing ZONE_DEVICE pages as the flows in question lack the underlying machinery for proper handling of ZONE_DEVICE pages.

This fixes a hang reported by Adam Borowski[*] in dev_pagemap_cleanup() when running a KVM guest backed with /dev/dax memory, as KVM straight up doesn't put any references to ZONE_DEVICE pages acquired by gup().

Note, Dan Williams proposed an alternative solution of doing put_page() on ZONE_DEVICE pages immediately after gup() in order to simplify the auditing needed to ensure is_zone_device_page() is called if and only if the backing device is pinned (via gup()). But that approach would break kvm_vcpu_{un}map() as KVM requires the page to be pinned from map() 'til unmap() when accessing guest memory, unlike KVM's secondary MMU, which coordinates with mmu_notifier invalidations to avoid creating stale page references, i.e. doesn't rely on pages being pinned.

Quoted 5 lines

[*] http://lkml.kernel.org/r/[email protected]

Quoted 1 line

Perhaps add:

Fixes 3565fce3a659 ("mm, x86: get_user_pages() for dax mappings")

...since that was the first kernel that broke KVM's assumption about which pfn types needed to have the reference count managed.

maintainer
#4 lore raw

Re: [PATCH v2 1/3] KVM: MMU: Do not treat ZONE_DEVICE pages as being reserved

Quoted 7 lines · Dan Williams
Quoted 5 lines

[*] http://lkml.kernel.org/r/[email protected]

Quoted 8 lines
Quoted 1 line

Perhaps add:

Fixes 3565fce3a659 ("mm, x86: get_user_pages() for dax mappings")

...since that was the first kernel that broke KVM's assumption about which pfn types needed to have the reference count managed.

Done, thanks!

Paolo

Acked-by on v2 1/3
#5 lore raw

Re: [PATCH v2 1/3] KVM: MMU: Do not treat ZONE_DEVICE pages as being reserved

In reply to Sean Christopherson
Quoted 32 lines

Am 11.11.2019 um 23:12 schrieb Sean Christopherson <[email protected]>:

Explicitly exempt ZONE_DEVICE pages from kvm_is_reserved_pfn() and instead manually handle ZONE_DEVICE on a case-by-case basis. For things like page refcounts, KVM needs to treat ZONE_DEVICE pages like normal pages, e.g. put pages grabbed via gup(). But for flows such as setting A/D bits or shifting refcounts for transparent huge pages, KVM needs to to avoid processing ZONE_DEVICE pages as the flows in question lack the underlying machinery for proper handling of ZONE_DEVICE pages.

This fixes a hang reported by Adam Borowski[*] in dev_pagemap_cleanup() when running a KVM guest backed with /dev/dax memory, as KVM straight up doesn't put any references to ZONE_DEVICE pages acquired by gup().

Note, Dan Williams proposed an alternative solution of doing put_page() on ZONE_DEVICE pages immediately after gup() in order to simplify the auditing needed to ensure is_zone_device_page() is called if and only if the backing device is pinned (via gup()). But that approach would break kvm_vcpu_{un}map() as KVM requires the page to be pinned from map() 'til unmap() when accessing guest memory, unlike KVM's secondary MMU, which coordinates with mmu_notifier invalidations to avoid creating stale page references, i.e. doesn't rely on pages being pinned.

[*] http://lkml.kernel.org/r/[email protected]

Quoted 6 lines

---

arch/x86/kvm/mmu.c       |  8 ++++----
include/linux/kvm_host.h |  1 +
virt/kvm/kvm_main.c      | 26 +++++++++++++++++++++++---
3 files changed, 28 insertions(+), 7 deletions(-)

Thanks for taking care of this! Other KVM related code (PPC, vfio) also has a reserved check (see my series), I didn‘t have a look yet at the details, how reserved pages are treated. Will do so in the next weeks, after hollidays.

Quoted 95 lines
arch/x86/kvm/mmu.c +4 -4
index 24c23c66b226..bf82b1f2e834 100644
@@ -3306,7 +3306,7 @@ static void transparent_hugepage_adjust(struct kvm_vcpu *vcpu,
    * here.
    */
   if (!is_error_noslot_pfn(pfn) && !kvm_is_reserved_pfn(pfn) &&
-        level == PT_PAGE_TABLE_LEVEL &&
+        !kvm_is_zone_device_pfn(pfn) && level == PT_PAGE_TABLE_LEVEL &&
       PageTransCompoundMap(pfn_to_page(pfn)) &&
       !mmu_gfn_lpage_is_disallowed(vcpu, gfn, PT_DIRECTORY_LEVEL)) {
       unsigned long mask;
@@ -5914,9 +5914,9 @@ static bool kvm_mmu_zap_collapsible_spte(struct kvm *kvm,
        * the guest, and the guest page table is using 4K page size
        * mapping if the indirect sp has level = 1.
        */
-        if (sp->role.direct &&
-            !kvm_is_reserved_pfn(pfn) &&
-            PageTransCompoundMap(pfn_to_page(pfn))) {
+        if (sp->role.direct && !kvm_is_reserved_pfn(pfn) &&
+            !kvm_is_zone_device_pfn(pfn) &&
+            PageTransCompoundMap(pfn_to_page(pfn))) {
           pte_list_remove(rmap_head, sptep);
 
           if (kvm_available_flush_tlb_with_range())
include/linux/kvm_host.h +1 -0
index a817e446c9aa..4ad1cd7d2d4d 100644
@@ -966,6 +966,7 @@ int kvm_cpu_has_pending_timer(struct kvm_vcpu *vcpu);
 void kvm_vcpu_kick(struct kvm_vcpu *vcpu);
 
 bool kvm_is_reserved_pfn(kvm_pfn_t pfn);
+bool kvm_is_zone_device_pfn(kvm_pfn_t pfn);
 
 struct kvm_irq_ack_notifier {
   struct hlist_node link;
virt/kvm/kvm_main.c +23 -3
index b8534c6b8cf6..bc9d10a0a334 100644
@@ -149,10 +149,30 @@ __weak int kvm_arch_mmu_notifier_invalidate_range(struct kvm *kvm,
   return 0;
 }
 
+bool kvm_is_zone_device_pfn(kvm_pfn_t pfn)
+{
+    /*
+     * The metadata used by is_zone_device_page() to determine whether or
+     * not a page is ZONE_DEVICE is guaranteed to be valid if and only if
+     * the device has been pinned, e.g. by get_user_pages().  WARN if the
+     * page_count() is zero to help detect bad usage of this helper.
+     */
+    if (!pfn_valid(pfn) || WARN_ON_ONCE(!page_count(pfn_to_page(pfn))))
+        return false;
+
+    return is_zone_device_page(pfn_to_page(pfn));
+}
+
 bool kvm_is_reserved_pfn(kvm_pfn_t pfn)
 {
+    /*
+     * ZONE_DEVICE pages currently set PG_reserved, but from a refcounting
+     * perspective they are "normal" pages, albeit with slightly different
+     * usage rules.
+     */
   if (pfn_valid(pfn))
-        return PageReserved(pfn_to_page(pfn));
+        return PageReserved(pfn_to_page(pfn)) &&
+               !kvm_is_zone_device_pfn(pfn);
 
   return true;
 }
@@ -1865,7 +1885,7 @@ EXPORT_SYMBOL_GPL(kvm_release_pfn_dirty);
 
 void kvm_set_pfn_dirty(kvm_pfn_t pfn)
 {
-    if (!kvm_is_reserved_pfn(pfn)) {
+    if (!kvm_is_reserved_pfn(pfn) && !kvm_is_zone_device_pfn(pfn)) {
       struct page *page = pfn_to_page(pfn);
 
       SetPageDirty(page);
@@ -1875,7 +1895,7 @@ EXPORT_SYMBOL_GPL(kvm_set_pfn_dirty);
 
 void kvm_set_pfn_accessed(kvm_pfn_t pfn)
 {
-    if (!kvm_is_reserved_pfn(pfn))
+    if (!kvm_is_reserved_pfn(pfn) && !kvm_is_zone_device_pfn(pfn))
       mark_page_accessed(pfn_to_page(pfn));
 }
 EXPORT_SYMBOL_GPL(kvm_set_pfn_accessed);
Signature · 1 line
2.24.0
#6 lore raw

PATCH v2 2/3 KVM: x86/mmu: Remove superfluous is_error_pfn() check from THP adjust

In reply to Sean Christopherson

Replace the is_error_noslot_pfn() check in transparent_hugepage_adjust() with an is_noslot_pfn() check. thp_adjust() cannot be reached with an error pfn as it is always called after handle_abnormal_pfn(), which aborts the page fault handler if an error pfn is encountered. Don't bother future proofing thp_adjust() with a WARN on is_error_pfn(), as calling thp_adjust() before handle_abnormal_pfn() is impossible for all intents and purposes, e.g. thp_adjust() relies on being called after mmu_notifier_retry() and while holding mmu_lock, thus moving it would essentially require a complete rewrite of KVM's page fault handlers.

No functional change intended.

---
 arch/x86/kvm/mmu.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
arch/x86/kvm/mmu.c +1 -1
index bf82b1f2e834..c35c6fb2635a 100644
@@ -3305,7 +3305,7 @@ static void transparent_hugepage_adjust(struct kvm_vcpu *vcpu,
 	 * PT_PAGE_TABLE_LEVEL and there would be no adjustment done
 	 * here.
 	 */
-	if (!is_error_noslot_pfn(pfn) && !kvm_is_reserved_pfn(pfn) &&
+	if (!is_noslot_pfn(pfn) && !kvm_is_reserved_pfn(pfn) &&
 	    !kvm_is_zone_device_pfn(pfn) && level == PT_PAGE_TABLE_LEVEL &&
 	    PageTransCompoundMap(pfn_to_page(pfn)) &&
 	    !mmu_gfn_lpage_is_disallowed(vcpu, gfn, PT_DIRECTORY_LEVEL)) {
Signature · 1 line
2.24.0
#7 lore raw

PATCH v2 3/3 KVM: x86/mmu: Add helper to consolidate huge page promotion

In reply to Sean Christopherson

Add a helper, kvm_is_hugepage_allowed(), to consolidate identical code across transparent_hugepage_adjust() and kvm_mmu_zap_collapsible_spte().

No functional change intended.

---
 arch/x86/kvm/mmu.c | 15 +++++++++------
 1 file changed, 9 insertions(+), 6 deletions(-)
arch/x86/kvm/mmu.c +9 -6
index c35c6fb2635a..95163f3abb24 100644
@@ -3292,6 +3292,12 @@ static int kvm_handle_bad_page(struct kvm_vcpu *vcpu, gfn_t gfn, kvm_pfn_t pfn)
 	return -EFAULT;
 }
 
+static bool kvm_is_hugepage_allowed(kvm_pfn_t pfn)
+{
+	return !kvm_is_reserved_pfn(pfn) && !kvm_is_zone_device_pfn(pfn) &&
+	       PageTransCompoundMap(pfn_to_page(pfn));
+}
+
 static void transparent_hugepage_adjust(struct kvm_vcpu *vcpu,
 					gfn_t gfn, kvm_pfn_t *pfnp,
 					int *levelp)
@@ -3305,9 +3311,8 @@ static void transparent_hugepage_adjust(struct kvm_vcpu *vcpu,
 	 * PT_PAGE_TABLE_LEVEL and there would be no adjustment done
 	 * here.
 	 */
-	if (!is_noslot_pfn(pfn) && !kvm_is_reserved_pfn(pfn) &&
-	    !kvm_is_zone_device_pfn(pfn) && level == PT_PAGE_TABLE_LEVEL &&
-	    PageTransCompoundMap(pfn_to_page(pfn)) &&
+	if (!is_noslot_pfn(pfn) && level == PT_PAGE_TABLE_LEVEL &&
+	    kvm_is_hugepage_allowed(pfn) &&
 	    !mmu_gfn_lpage_is_disallowed(vcpu, gfn, PT_DIRECTORY_LEVEL)) {
 		unsigned long mask;
 		/*
@@ -5914,9 +5919,7 @@ static bool kvm_mmu_zap_collapsible_spte(struct kvm *kvm,
 		 * the guest, and the guest page table is using 4K page size
 		 * mapping if the indirect sp has level = 1.
 		 */
-		if (sp->role.direct && !kvm_is_reserved_pfn(pfn) &&
-		    !kvm_is_zone_device_pfn(pfn) &&
-		    PageTransCompoundMap(pfn_to_page(pfn))) {
+		if (sp->role.direct && kvm_is_hugepage_allowed(pfn)) {
 			pte_list_remove(rmap_head, sptep);
 
 			if (kvm_available_flush_tlb_with_range())
Signature · 1 line
2.24.0