Skip to contentSkip to conversation

KVM: x86/mmu: Add macro for hugepage GFN mask

PATCH In review Sean Christopherson 3 participants 12 messages Oct 27, 2020 – Nov 5, 2020 1 message not in the archive

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

cover KVM: x86/mmu: Add macro for hugepage GFN mask

Add a macro, which is probably long overdue, to replace open coded variants of "~(KVM_PAGES_PER_HPAGE(level) - 1)". The straw that broke the camel's back is the TDP MMU's round_gfn_for_level(), which goes straight for the optimized approach of using NEG instead of SUB+NOT (gcc uses NEG for both). The use of '-(...)' made me do a double take (more like a quadrupal take) when reading the TDP MMU code as my eyes/brain have been heavily trained to look for the more common '~(... - 1)'.

Sean Christopherson (3):
KVM: x86/mmu: Add helper macro for computing hugepage GFN mask
KVM: x86/mmu: Open code GFN "rounding" in TDP MMU
KVM: x86/mmu: Use hugepage GFN mask to compute GFN offset mask

 arch/x86/include/asm/kvm_host.h |  1 +
 arch/x86/kvm/mmu/mmu.c          |  4 ++--
 arch/x86/kvm/mmu/mmutrace.h     |  2 +-
 arch/x86/kvm/mmu/paging_tmpl.h  |  4 ++--
 arch/x86/kvm/mmu/tdp_iter.c     | 11 +++--------
 arch/x86/kvm/mmu/tdp_mmu.c      |  2 +-
 arch/x86/kvm/x86.c              |  6 +++---
 7 files changed, 13 insertions(+), 17 deletions(-)
Signature · 1 line
2.28.0
#1 lore raw

PATCH 1/3 KVM: x86/mmu: Add helper macro for computing hugepage GFN mask

Add a helper to compute the GFN mask given a hugepage level, KVM is accumulating quite a few users with the addition of the TDP MMU.

Note, gcc is clever enough to use a single NEG instruction instead of SUB+NOT, i.e. use the more common "~(level -1)" pattern instead of round_gfn_for_level()'s direct two's complement trickery.

---
 arch/x86/include/asm/kvm_host.h | 1 +
 arch/x86/kvm/mmu/mmu.c          | 2 +-
 arch/x86/kvm/mmu/paging_tmpl.h  | 4 ++--
 arch/x86/kvm/mmu/tdp_iter.c     | 2 +-
 4 files changed, 5 insertions(+), 4 deletions(-)
arch/x86/include/asm/kvm_host.h +1 -0
index d44858b69353..6ea046415f29 100644
@@ -119,6 +119,7 @@
 #define KVM_HPAGE_SIZE(x)	(1UL << KVM_HPAGE_SHIFT(x))
 #define KVM_HPAGE_MASK(x)	(~(KVM_HPAGE_SIZE(x) - 1))
 #define KVM_PAGES_PER_HPAGE(x)	(KVM_HPAGE_SIZE(x) / PAGE_SIZE)
+#define KVM_HPAGE_GFN_MASK(x)	(~(KVM_PAGES_PER_HPAGE(x) - 1))
 
 static inline gfn_t gfn_to_index(gfn_t gfn, gfn_t base_gfn, int level)
 {
arch/x86/kvm/mmu/mmu.c +1 -1
index 17587f496ec7..3bfc7ee44e51 100644
@@ -2886,7 +2886,7 @@ static int __direct_map(struct kvm_vcpu *vcpu, gpa_t gpa, u32 error_code,
 			disallowed_hugepage_adjust(*it.sptep, gfn, it.level,
 						   &pfn, &level);
 
-		base_gfn = gfn & ~(KVM_PAGES_PER_HPAGE(it.level) - 1);
+		base_gfn = gfn & KVM_HPAGE_GFN_MASK(it.level);
 		if (it.level == level)
 			break;
 
arch/x86/kvm/mmu/paging_tmpl.h +2 -2
index 50e268eb8e1a..76ee36f2afd2 100644
@@ -698,7 +698,7 @@ static int FNAME(fetch)(struct kvm_vcpu *vcpu, gpa_t addr,
 			disallowed_hugepage_adjust(*it.sptep, gw->gfn, it.level,
 						   &pfn, &level);
 
-		base_gfn = gw->gfn & ~(KVM_PAGES_PER_HPAGE(it.level) - 1);
+		base_gfn = gw->gfn & KVM_HPAGE_GFN_MASK(it.level);
 		if (it.level == level)
 			break;
 
@@ -751,7 +751,7 @@ FNAME(is_self_change_mapping)(struct kvm_vcpu *vcpu,
 			      bool *write_fault_to_shadow_pgtable)
 {
 	int level;
-	gfn_t mask = ~(KVM_PAGES_PER_HPAGE(walker->level) - 1);
+	gfn_t mask = KVM_HPAGE_GFN_MASK(walker->level);
 	bool self_changed = false;
 
 	if (!(walker->pte_access & ACC_WRITE_MASK ||
arch/x86/kvm/mmu/tdp_iter.c +1 -1
index 87b7e16911db..c6e914c96641 100644
@@ -17,7 +17,7 @@ static void tdp_iter_refresh_sptep(struct tdp_iter *iter)
 
 static gfn_t round_gfn_for_level(gfn_t gfn, int level)
 {
-	return gfn & -KVM_PAGES_PER_HPAGE(level);
+	return gfn & KVM_HPAGE_GFN_MASK(level);
 }
 
 /*
Signature · 1 line
2.28.0
Reviewed-by on v1 1/3
#2 lore raw

Re: [PATCH 1/3] KVM: x86/mmu: Add helper macro for computing hugepage GFN mask

Quoted 7 lines · Sean Christopherson

Add a helper to compute the GFN mask given a hugepage level, KVM is accumulating quite a few users with the addition of the TDP MMU.

Note, gcc is clever enough to use a single NEG instruction instead of SUB+NOT, i.e. use the more common "~(level -1)" pattern instead of round_gfn_for_level()'s direct two's complement trickery.

As far as I can tell this patch has no functional changes intended. Please correct me if that's not correct.

Quoted 2 lines
Quoted 12 lines
---
 arch/x86/include/asm/kvm_host.h | 1 +
 arch/x86/kvm/mmu/mmu.c          | 2 +-
 arch/x86/kvm/mmu/paging_tmpl.h  | 4 ++--
 arch/x86/kvm/mmu/tdp_iter.c     | 2 +-
 4 files changed, 5 insertions(+), 4 deletions(-)
arch/x86/include/asm/kvm_host.h +0 -0
index d44858b69353..6ea046415f29 100644

arch/x86/include/asm/kvm_host.h line 122

Quoted 4 lines
arch/x86/include/asm/kvm_host.h +1 -0
@@ -119,6 +119,7 @@
 #define KVM_HPAGE_SIZE(x)      (1UL << KVM_HPAGE_SHIFT(x))
 #define KVM_HPAGE_MASK(x)      (~(KVM_HPAGE_SIZE(x) - 1))
 #define KVM_PAGES_PER_HPAGE(x) (KVM_HPAGE_SIZE(x) / PAGE_SIZE)
+#define KVM_HPAGE_GFN_MASK(x)  (~(KVM_PAGES_PER_HPAGE(x) - 1))

NIT: I know x follows the convention on adjacent macros, but this would be clearer to me if x was changed to level. (Probably for all the macros in this block)

Quoted 54 lines
arch/x86/include/asm/kvm_host.h +0 -0
 static inline gfn_t gfn_to_index(gfn_t gfn, gfn_t base_gfn, int level)
 {
arch/x86/kvm/mmu/mmu.c +1 -1
index 17587f496ec7..3bfc7ee44e51 100644
@@ -2886,7 +2886,7 @@ static int __direct_map(struct kvm_vcpu *vcpu, gpa_t gpa, u32 error_code,
                        disallowed_hugepage_adjust(*it.sptep, gfn, it.level,
                                                   &pfn, &level);
 
-               base_gfn = gfn & ~(KVM_PAGES_PER_HPAGE(it.level) - 1);
+               base_gfn = gfn & KVM_HPAGE_GFN_MASK(it.level);
                if (it.level == level)
                        break;
 
arch/x86/kvm/mmu/paging_tmpl.h +2 -2
index 50e268eb8e1a..76ee36f2afd2 100644
@@ -698,7 +698,7 @@ static int FNAME(fetch)(struct kvm_vcpu *vcpu, gpa_t addr,
                        disallowed_hugepage_adjust(*it.sptep, gw->gfn, it.level,
                                                   &pfn, &level);
 
-               base_gfn = gw->gfn & ~(KVM_PAGES_PER_HPAGE(it.level) - 1);
+               base_gfn = gw->gfn & KVM_HPAGE_GFN_MASK(it.level);
                if (it.level == level)
                        break;
 
@@ -751,7 +751,7 @@ FNAME(is_self_change_mapping)(struct kvm_vcpu *vcpu,
                              bool *write_fault_to_shadow_pgtable)
 {
        int level;
-       gfn_t mask = ~(KVM_PAGES_PER_HPAGE(walker->level) - 1);
+       gfn_t mask = KVM_HPAGE_GFN_MASK(walker->level);
        bool self_changed = false;
 
        if (!(walker->pte_access & ACC_WRITE_MASK ||
arch/x86/kvm/mmu/tdp_iter.c +1 -2
index 87b7e16911db..c6e914c96641 100644
@@ -17,7 +17,7 @@ static void tdp_iter_refresh_sptep(struct tdp_iter *iter)
 
 static gfn_t round_gfn_for_level(gfn_t gfn, int level)
 {
-       return gfn & -KVM_PAGES_PER_HPAGE(level);
+       return gfn & KVM_HPAGE_GFN_MASK(level);
 }
 
 /*
--

2.28.0

#3 lore raw

Re: [PATCH 1/3] KVM: x86/mmu: Add helper macro for computing hugepage GFN mask

Quoted 6 lines · Ben Gardon
Quoted 4 lines · Sean Christopherson

Add a helper to compute the GFN mask given a hugepage level, KVM is accumulating quite a few users with the addition of the TDP MMU.

Quoted 6 lines
Quoted 3 lines

Note, gcc is clever enough to use a single NEG instruction instead of SUB+NOT, i.e. use the more common "~(level -1)" pattern instead of round_gfn_for_level()'s direct two's complement trickery.

As far as I can tell this patch has no functional changes intended. Please correct me if that's not correct.

Correct. :-)

Quoted 12 lines
Quoted 2 lines
Quoted 7 lines
---
 arch/x86/include/asm/kvm_host.h | 1 +
 arch/x86/kvm/mmu/mmu.c          | 2 +-
 arch/x86/kvm/mmu/paging_tmpl.h  | 4 ++--
 arch/x86/kvm/mmu/tdp_iter.c     | 2 +-
 4 files changed, 5 insertions(+), 4 deletions(-)
Quoted 13 lines

arch/x86/include/asm/kvm_host.h line 122

Quoted 9 lines
arch/x86/include/asm/kvm_host.h +1 -0
index d44858b69353..6ea046415f29 100644
@@ -119,6 +119,7 @@
 #define KVM_HPAGE_SIZE(x)      (1UL << KVM_HPAGE_SHIFT(x))
 #define KVM_HPAGE_MASK(x)      (~(KVM_HPAGE_SIZE(x) - 1))
 #define KVM_PAGES_PER_HPAGE(x) (KVM_HPAGE_SIZE(x) / PAGE_SIZE)
+#define KVM_HPAGE_GFN_MASK(x)  (~(KVM_PAGES_PER_HPAGE(x) - 1))

NIT: I know x follows the convention on adjacent macros, but this would be clearer to me if x was changed to level. (Probably for all the macros in this block)

Agreed. I'll spin a v2 and opportunistically change them all to "level" in this patch. I'll also add "No function change intended™." to patches 1 and 3.

Quoted 2 lines
Quoted 2 lines
arch/x86/include/asm/kvm_host.h +0 -0
 static inline gfn_t gfn_to_index(gfn_t gfn, gfn_t base_gfn, int level)
 {
#4 lore raw

PATCH 2/3 KVM: x86/mmu: Open code GFN "rounding" in TDP MMU

In reply to Sean Christopherson

Drop round_gfn_for_level() and directly use the recently introdocued KVM_HPAGE_GFN_MASK() macro. Hiding the masking in a "rounding" function adds an extra "what does this do?" lookup, whereas the concept and usage of PFN/GFN masks is common enough that it's easy to read the open coded version without thinking too hard.

No functional change intended.

---
 arch/x86/kvm/mmu/tdp_iter.c | 11 +++--------
 1 file changed, 3 insertions(+), 8 deletions(-)
arch/x86/kvm/mmu/tdp_iter.c +3 -8
index c6e914c96641..4175947dc401 100644
@@ -15,11 +15,6 @@ static void tdp_iter_refresh_sptep(struct tdp_iter *iter)
 	iter->old_spte = READ_ONCE(*iter->sptep);
 }
 
-static gfn_t round_gfn_for_level(gfn_t gfn, int level)
-{
-	return gfn & KVM_HPAGE_GFN_MASK(level);
-}
-
 /*
  * Sets a TDP iterator to walk a pre-order traversal of the paging structure
  * rooted at root_pt, starting with the walk to translate goal_gfn.
@@ -36,7 +31,7 @@ void tdp_iter_start(struct tdp_iter *iter, u64 *root_pt, int root_level,
 	iter->level = root_level;
 	iter->pt_path[iter->level - 1] = root_pt;
 
-	iter->gfn = round_gfn_for_level(iter->goal_gfn, iter->level);
+	iter->gfn = iter->goal_gfn & KVM_HPAGE_GFN_MASK(iter->level);
 	tdp_iter_refresh_sptep(iter);
 
 	iter->valid = true;
@@ -82,7 +77,7 @@ static bool try_step_down(struct tdp_iter *iter)
 
 	iter->level--;
 	iter->pt_path[iter->level - 1] = child_pt;
-	iter->gfn = round_gfn_for_level(iter->goal_gfn, iter->level);
+	iter->gfn = iter->goal_gfn & KVM_HPAGE_GFN_MASK(iter->level);
 	tdp_iter_refresh_sptep(iter);
 
 	return true;
@@ -124,7 +119,7 @@ static bool try_step_up(struct tdp_iter *iter)
 		return false;
 
 	iter->level++;
-	iter->gfn = round_gfn_for_level(iter->gfn, iter->level);
+	iter->gfn &= KVM_HPAGE_GFN_MASK(iter->level);
 	tdp_iter_refresh_sptep(iter);
 
 	return true;
Signature · 1 line
2.28.0
Reviewed-by on v1 2/3
#5 lore raw

Re: [PATCH 2/3] KVM: x86/mmu: Open code GFN "rounding" in TDP MMU

Quoted 11 lines · Sean Christopherson

Drop round_gfn_for_level() and directly use the recently introdocued KVM_HPAGE_GFN_MASK() macro. Hiding the masking in a "rounding" function adds an extra "what does this do?" lookup, whereas the concept and usage of PFN/GFN masks is common enough that it's easy to read the open coded version without thinking too hard.

No functional change intended.

Quoted 50 lines
---
 arch/x86/kvm/mmu/tdp_iter.c | 11 +++--------
 1 file changed, 3 insertions(+), 8 deletions(-)
arch/x86/kvm/mmu/tdp_iter.c +3 -9
index c6e914c96641..4175947dc401 100644
@@ -15,11 +15,6 @@ static void tdp_iter_refresh_sptep(struct tdp_iter *iter)
        iter->old_spte = READ_ONCE(*iter->sptep);
 }
 
-static gfn_t round_gfn_for_level(gfn_t gfn, int level)
-{
-       return gfn & KVM_HPAGE_GFN_MASK(level);
-}
-
 /*
  * Sets a TDP iterator to walk a pre-order traversal of the paging structure
  * rooted at root_pt, starting with the walk to translate goal_gfn.
@@ -36,7 +31,7 @@ void tdp_iter_start(struct tdp_iter *iter, u64 *root_pt, int root_level,
        iter->level = root_level;
        iter->pt_path[iter->level - 1] = root_pt;
 
-       iter->gfn = round_gfn_for_level(iter->goal_gfn, iter->level);
+       iter->gfn = iter->goal_gfn & KVM_HPAGE_GFN_MASK(iter->level);
        tdp_iter_refresh_sptep(iter);
 
        iter->valid = true;
@@ -82,7 +77,7 @@ static bool try_step_down(struct tdp_iter *iter)
 
        iter->level--;
        iter->pt_path[iter->level - 1] = child_pt;
-       iter->gfn = round_gfn_for_level(iter->goal_gfn, iter->level);
+       iter->gfn = iter->goal_gfn & KVM_HPAGE_GFN_MASK(iter->level);
        tdp_iter_refresh_sptep(iter);
 
        return true;
@@ -124,7 +119,7 @@ static bool try_step_up(struct tdp_iter *iter)
                return false;
 
        iter->level++;
-       iter->gfn = round_gfn_for_level(iter->gfn, iter->level);
+       iter->gfn &= KVM_HPAGE_GFN_MASK(iter->level);
        tdp_iter_refresh_sptep(iter);
 
        return true;
--

2.28.0

#6 lore raw

PATCH 3/3 KVM: x86/mmu: Use hugepage GFN mask to compute GFN offset mask

In reply to Sean Christopherson

Use the logical NOT of KVM_HPAGE_GFN_MASK() to compute the GFN offset mask instead of open coding the equivalent in a variety of locations.

---
 arch/x86/kvm/mmu/mmu.c      | 2 +-
 arch/x86/kvm/mmu/mmutrace.h | 2 +-
 arch/x86/kvm/mmu/tdp_mmu.c  | 2 +-
 arch/x86/kvm/x86.c          | 6 +++---
 4 files changed, 6 insertions(+), 6 deletions(-)
arch/x86/kvm/mmu/mmu.c +1 -1
index 3bfc7ee44e51..9fb50c666ec5 100644
@@ -2827,7 +2827,7 @@ int kvm_mmu_hugepage_adjust(struct kvm_vcpu *vcpu, gfn_t gfn,
 	 * mmu_notifier_retry() was successful and mmu_lock is held, so
 	 * the pmd can't be split from under us.
 	 */
-	mask = KVM_PAGES_PER_HPAGE(level) - 1;
+	mask = ~KVM_HPAGE_GFN_MASK(level);
 	VM_BUG_ON((gfn & mask) != (pfn & mask));
 	*pfnp = pfn & ~mask;
 
arch/x86/kvm/mmu/mmutrace.h +1 -1
index 213699b27b44..4432ca3c7e4e 100644
@@ -372,7 +372,7 @@ TRACE_EVENT(
 
 	TP_fast_assign(
 		__entry->gfn = addr >> PAGE_SHIFT;
-		__entry->pfn = pfn | (__entry->gfn & (KVM_PAGES_PER_HPAGE(level) - 1));
+		__entry->pfn = pfn | (__entry->gfn & ~KVM_HPAGE_GFN_MASK(level));
 		__entry->level = level;
 	),
 
arch/x86/kvm/mmu/tdp_mmu.c +1 -1
index 27e381c9da6c..681686608c0b 100644
@@ -209,7 +209,7 @@ static void __handle_changed_spte(struct kvm *kvm, int as_id, gfn_t gfn,
 
 	WARN_ON(level > PT64_ROOT_MAX_LEVEL);
 	WARN_ON(level < PG_LEVEL_4K);
-	WARN_ON(gfn & (KVM_PAGES_PER_HPAGE(level) - 1));
+	WARN_ON(gfn & ~KVM_HPAGE_GFN_MASK(level));
 
 	/*
 	 * If this warning were to trigger it would indicate that there was a
arch/x86/kvm/x86.c +3 -3
index 397f599b20e5..faf4c4ddde94 100644
@@ -10451,16 +10451,16 @@ static int kvm_alloc_memslot_metadata(struct kvm_memory_slot *slot,
 
 		slot->arch.lpage_info[i - 1] = linfo;
 
-		if (slot->base_gfn & (KVM_PAGES_PER_HPAGE(level) - 1))
+		if (slot->base_gfn & ~KVM_HPAGE_GFN_MASK(level))
 			linfo[0].disallow_lpage = 1;
-		if ((slot->base_gfn + npages) & (KVM_PAGES_PER_HPAGE(level) - 1))
+		if ((slot->base_gfn + npages) & ~KVM_HPAGE_GFN_MASK(level))
 			linfo[lpages - 1].disallow_lpage = 1;
 		ugfn = slot->userspace_addr >> PAGE_SHIFT;
 		/*
 		 * If the gfn and userspace address are not aligned wrt each
 		 * other, disable large page support for this slot.
 		 */
-		if ((slot->base_gfn ^ ugfn) & (KVM_PAGES_PER_HPAGE(level) - 1)) {
+		if ((slot->base_gfn ^ ugfn) & ~KVM_HPAGE_GFN_MASK(level)) {
 			unsigned long j;
 
 			for (j = 0; j < lpages; ++j)
Signature · 1 line
2.28.0
Reviewed-by on v1 3/3
#7 lore raw

Re: [PATCH 3/3] KVM: x86/mmu: Use hugepage GFN mask to compute GFN offset mask

Quoted 3 lines · Sean Christopherson

Use the logical NOT of KVM_HPAGE_GFN_MASK() to compute the GFN offset mask instead of open coding the equivalent in a variety of locations.

I don't see a "no functional change expected" note on this patch as was on the previous one, but I don't think this represents any functional change.

Quoted 2 lines
Quoted 73 lines
---
 arch/x86/kvm/mmu/mmu.c      | 2 +-
 arch/x86/kvm/mmu/mmutrace.h | 2 +-
 arch/x86/kvm/mmu/tdp_mmu.c  | 2 +-
 arch/x86/kvm/x86.c          | 6 +++---
 4 files changed, 6 insertions(+), 6 deletions(-)
arch/x86/kvm/mmu/mmu.c +1 -1
index 3bfc7ee44e51..9fb50c666ec5 100644
@@ -2827,7 +2827,7 @@ int kvm_mmu_hugepage_adjust(struct kvm_vcpu *vcpu, gfn_t gfn,
         * mmu_notifier_retry() was successful and mmu_lock is held, so
         * the pmd can't be split from under us.
         */
-       mask = KVM_PAGES_PER_HPAGE(level) - 1;
+       mask = ~KVM_HPAGE_GFN_MASK(level);
        VM_BUG_ON((gfn & mask) != (pfn & mask));
        *pfnp = pfn & ~mask;
 
arch/x86/kvm/mmu/mmutrace.h +1 -1
index 213699b27b44..4432ca3c7e4e 100644
@@ -372,7 +372,7 @@ TRACE_EVENT(
 
        TP_fast_assign(
                __entry->gfn = addr >> PAGE_SHIFT;
-               __entry->pfn = pfn | (__entry->gfn & (KVM_PAGES_PER_HPAGE(level) - 1));
+               __entry->pfn = pfn | (__entry->gfn & ~KVM_HPAGE_GFN_MASK(level));
                __entry->level = level;
        ),
 
arch/x86/kvm/mmu/tdp_mmu.c +1 -1
index 27e381c9da6c..681686608c0b 100644
@@ -209,7 +209,7 @@ static void __handle_changed_spte(struct kvm *kvm, int as_id, gfn_t gfn,
 
        WARN_ON(level > PT64_ROOT_MAX_LEVEL);
        WARN_ON(level < PG_LEVEL_4K);
-       WARN_ON(gfn & (KVM_PAGES_PER_HPAGE(level) - 1));
+       WARN_ON(gfn & ~KVM_HPAGE_GFN_MASK(level));
 
        /*
         * If this warning were to trigger it would indicate that there was a
arch/x86/kvm/x86.c +3 -4
index 397f599b20e5..faf4c4ddde94 100644
@@ -10451,16 +10451,16 @@ static int kvm_alloc_memslot_metadata(struct kvm_memory_slot *slot,
 
                slot->arch.lpage_info[i - 1] = linfo;
 
-               if (slot->base_gfn & (KVM_PAGES_PER_HPAGE(level) - 1))
+               if (slot->base_gfn & ~KVM_HPAGE_GFN_MASK(level))
                        linfo[0].disallow_lpage = 1;
-               if ((slot->base_gfn + npages) & (KVM_PAGES_PER_HPAGE(level) - 1))
+               if ((slot->base_gfn + npages) & ~KVM_HPAGE_GFN_MASK(level))
                        linfo[lpages - 1].disallow_lpage = 1;
                ugfn = slot->userspace_addr >> PAGE_SHIFT;
                /*
                 * If the gfn and userspace address are not aligned wrt each
                 * other, disable large page support for this slot.
                 */
-               if ((slot->base_gfn ^ ugfn) & (KVM_PAGES_PER_HPAGE(level) - 1)) {
+               if ((slot->base_gfn ^ ugfn) & ~KVM_HPAGE_GFN_MASK(level)) {
                        unsigned long j;
 
                        for (j = 0; j < lpages; ++j)
--

2.28.0

#8 lore raw

Re: [PATCH 3/3] KVM: x86/mmu: Use hugepage GFN mask to compute GFN offset mask

Quoted 9 lines · Ben Gardon
Quoted 3 lines · Sean Christopherson

Use the logical NOT of KVM_HPAGE_GFN_MASK() to compute the GFN offset mask instead of open coding the equivalent in a variety of locations.

I don't see a "no functional change expected" note on this patch as was on the previous one, but I don't think this represents any functional change.

Ah, yeah, I meant to call out in the cover letter than nothing in this series generates a functional difference, e.g. objdump of kvm/kvm-intel is identical from start to finish.

maintainer
#9 lore raw
In reply to Sean Christopherson
Quoted 7 lines · Sean Christopherson

Add a macro, which is probably long overdue, to replace open coded variants of "~(KVM_PAGES_PER_HPAGE(level) - 1)". The straw that broke the camel's back is the TDP MMU's round_gfn_for_level(), which goes straight for the optimized approach of using NEG instead of SUB+NOT (gcc uses NEG for both). The use of '-(...)' made me do a double take (more like a quadrupal take) when reading the TDP MMU code as my eyes/brain have been heavily trained to look for the more common '~(... - 1)'.

The upside is that a "how many" macro such as KVM_PAGES_PER_HPAGE will have only one definition that makes sense. With a "mask" macro you never know if the 1s are to the left or to the right. That is, does "gfn & KVM_HPAGE_GFN_MASK(x)" return the first gfn within the huge page or the index of the page within the huge page?

This is actually a problem with this series; see this line in patch 3:

- mask = KVM_PAGES_PER_HPAGE(level) - 1;

+	mask = ~KVM_HPAGE_GFN_MASK(level);
        ^^^^                  ^^^^

So it's a mask, but not _that_ mask, _another_ mask. :) That's why I don't really like "mask" macros, now the other question is how to express bit masking with "how many" macros.

First of all, I think we all agree that when reading "x & how_many()" we assume (or we check first thing of all) that the right side is a power of two. I like "x & -y" because---even if your eyes have trouble distinguishing "-" from "~"---it's clearly not "x & (y-1)" and therefore the only sensible operation that the AND can do "clear everything to the right".

Now I realize that maybe my obsession for bit twiddling tricks is not shared with everyone, and of course if you're debugging it you have to look closer and check if it's really "x & -y" or "x & ~y", but at least in normal cursory code reading that's how it works for me.

Paolo

A message that is not in the archive (sent off-list or not yet received): [email protected]

maintainer
#11 lore raw
Quoted 11 lines · Sean Christopherson

The naming and usage also aligns with the kernel, which defines PAGE, PMD and PUD masks, and has near identical usage patterns.

  #define PAGE_SIZE               (_AC(1,UL) << PAGE_SHIFT)
  #define PAGE_MASK               (~(PAGE_SIZE-1))

  #define PMD_PAGE_SIZE           (_AC(1, UL) << PMD_SHIFT)
  #define PMD_PAGE_MASK           (~(PMD_PAGE_SIZE-1))

  #define PUD_PAGE_SIZE           (_AC(1, UL) << PUD_SHIFT)
  #define PUD_PAGE_MASK           (~(PUD_PAGE_SIZE-1))

Well, PAGE_MASK is also one of my pet peeves for Linux. At least I am consistent. :)

Quoted 7 lines
Quoted 3 lines

and of course if you're debugging it you have to
look closer and check if it's really "x & -y" or "x & ~y", but at least in normal cursory code reading that's how it works for me.

IMO, "x & -y" has a higher barrier to entry, especially when the kernel's page masks uses "x & ~(y - 1))". But, my opinion is definitely colored by my inability to read two's-complement on the fly.

Fair enough. What about having instead

#define KVM_HPAGE_GFN_BASE(gfn, level) \
(x & ~(KVM_PAGES_PER_HPAGE(gfn) - 1))
#define KVM_HPAGE_GFN_INDEX(gfn, level) \
(x & (KVM_PAGES_PER_HPAGE(gfn) - 1))

?

Paolo

#12 lore raw
Quoted 24 lines · Paolo Bonzini
Quoted 11 lines · Sean Christopherson

The naming and usage also aligns with the kernel, which defines PAGE, PMD and PUD masks, and has near identical usage patterns.

  #define PAGE_SIZE               (_AC(1,UL) << PAGE_SHIFT)
  #define PAGE_MASK               (~(PAGE_SIZE-1))

  #define PMD_PAGE_SIZE           (_AC(1, UL) << PMD_SHIFT)
  #define PMD_PAGE_MASK           (~(PMD_PAGE_SIZE-1))

  #define PUD_PAGE_SIZE           (_AC(1, UL) << PUD_SHIFT)
  #define PUD_PAGE_MASK           (~(PUD_PAGE_SIZE-1))

Well, PAGE_MASK is also one of my pet peeves for Linux. At least I am consistent. :)

Quoted 7 lines
Quoted 3 lines

and of course if you're debugging it you have to
look closer and check if it's really "x & -y" or "x & ~y", but at least in normal cursory code reading that's how it works for me.

IMO, "x & -y" has a higher barrier to entry, especially when the kernel's page masks uses "x & ~(y - 1))". But, my opinion is definitely colored by my inability to read two's-complement on the fly.

Quoted 8 lines

Fair enough. What about having instead

#define KVM_HPAGE_GFN_BASE(gfn, level)  \
   (x & ~(KVM_PAGES_PER_HPAGE(gfn) - 1))
#define KVM_HPAGE_GFN_INDEX(gfn, level)  \
   (x & (KVM_PAGES_PER_HPAGE(gfn) - 1))

?

Hmm, not awful? What about OFFSET instead of INDEX, to pair with page offset? I don't particularly love either one, but I can't think of anything better.