Skip to contentSkip to conversation

KVM: Use kvm_pfn_t for local PFN variable in hva_to_pfn_remapped()

PATCH Applied Sean Christopherson 2 participants 2 messages Feb 8, 2021 – Feb 9, 2021 KVM

lore
Share & tools
lore

Sean Christopherson

author reviewer
#0 lore raw

PATCH 1/1 KVM: Use kvm_pfn_t for local PFN variable in hva_to_pfn_remapped()

Use kvm_pfn_t, a.k.a. u64, for the local 'pfn' variable when retrieving a so called "remapped" hva/pfn pair. In theory, the hva could resolve to a pfn in high memory on a 32-bit kernel.

This bug was inadvertantly exposed by commit bd2fae8da794 ("KVM: do not assume PTE is writable after follow_pfn"), which added an error PFN value to the mix, causing gcc to comlain about overflowing the unsigned long.

arch/x86/kvm/../../../virt/kvm/kvm_main.c: In function ‘hva_to_pfn_remapped’:
include/linux/kvm_host.h:89:30: error: conversion from ‘long long unsigned int’
to ‘long unsigned int’ changes value from ‘9218868437227405314’ to ‘2’ [-Werror=overflow]
89 | #define KVM_PFN_ERR_RO_FAULT (KVM_PFN_ERR_MASK + 2)
| ^
virt/kvm/kvm_main.c:1935:9: note: in expansion of macro ‘KVM_PFN_ERR_RO_FAULT’

Fixes add6a0cd1c5b ("KVM: MMU: try to fix up page faults before giving up")

---

I don't actually know that it's possible for a remapped pfn to be a 64-bit value on a stock kernel. But, backporting a one-liner is far easier and safer than trying to audit all possible flows. :-)

 virt/kvm/kvm_main.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
virt/kvm/kvm_main.c +1 -1
index ee4ac2618ec5..001b9de4e727 100644
@@ -1906,7 +1906,7 @@ static int hva_to_pfn_remapped(struct vm_area_struct *vma,
 			       bool write_fault, bool *writable,
 			       kvm_pfn_t *p_pfn)
 {
-	unsigned long pfn;
+	kvm_pfn_t pfn;
 	pte_t *ptep;
 	spinlock_t *ptl;
 	int r;
Signature · 1 line
2.30.0.478.g8a0d178c01-goog
maintainer Applied
#1 lore raw
Quoted 34 lines · Sean Christopherson

Use kvm_pfn_t, a.k.a. u64, for the local 'pfn' variable when retrieving a so called "remapped" hva/pfn pair. In theory, the hva could resolve to a pfn in high memory on a 32-bit kernel.

This bug was inadvertantly exposed by commit bd2fae8da794 ("KVM: do not assume PTE is writable after follow_pfn"), which added an error PFN value to the mix, causing gcc to comlain about overflowing the unsigned long.

   arch/x86/kvm/../../../virt/kvm/kvm_main.c: In function ‘hva_to_pfn_remapped’:
   include/linux/kvm_host.h:89:30: error: conversion from ‘long long unsigned int’
                                   to ‘long unsigned int’ changes value from
                                   ‘9218868437227405314’ to ‘2’ [-Werror=overflow]
    89 | #define KVM_PFN_ERR_RO_FAULT (KVM_PFN_ERR_MASK + 2)
       |                              ^

virt/kvm/kvm_main.c:1935:9: note: in expansion of macro ‘KVM_PFN_ERR_RO_FAULT’

Fixes add6a0cd1c5b ("KVM: MMU: try to fix up page faults before giving up")

---

I don't actually know that it's possible for a remapped pfn to be a 64-bit value on a stock kernel. But, backporting a one-liner is far easier and safer than trying to audit all possible flows. :-)

  virt/kvm/kvm_main.c | 2 +-
  1 file changed, 1 insertion(+), 1 deletion(-)
virt/kvm/kvm_main.c +0 -0
index ee4ac2618ec5..001b9de4e727 100644

virt/kvm/kvm_main.c static int hva_to_pfn_remapped(struct vm_area_struct *vma, · line 1912

Quoted 8 lines
virt/kvm/kvm_main.c +1 -1
@@ -1906,7 +1906,7 @@ static int hva_to_pfn_remapped(struct vm_area_struct *vma,
  			       bool write_fault, bool *writable,
  			       kvm_pfn_t *p_pfn)
  {
-	unsigned long pfn;
+	kvm_pfn_t pfn;
  	pte_t *ptep;
  	spinlock_t *ptl;
  	int r;

Queued, thanks.

Paolo