Skip to contentSkip to conversation

kvm: vmx: fix limit checking in get_vmx_mem_address()

PATCH In review Eugene Korenevsky 2 participants 2 messages Jun 7, 2019

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

PATCH v5 1/3 kvm: vmx: fix limit checking in get_vmx_mem_address()

Intel SDM vol. 3, 5.3:
The processor causes a
general-protection exception (or, if the segment is SS, a stack-fault exception) any time an attempt is made to access the following addresses in a segment:
- A byte at an offset greater than the effective limit
- A word at an offset greater than the (effective-limit – 1)
- A doubleword at an offset greater than the (effective-limit – 3)
- A quadword at an offset greater than the (effective-limit – 7)

Therefore, the generic limit checking error condition must be

exn = (off > limit + 1 - access_len) = (off + access_len - 1 > limit)

but not

exn = (off + access_len > limit)

as for now.

Also avoid integer overflow of off at 32-bit KVM by casting it to u64.

Note: access length is currently sizeof(u64) which is incorrect. This will be fixed in the subsequent patch.

---
Changes in v3 since v2: fixed limit checking condition to avoid underflow; added note
Changes in v4 since v3: fixed off overflow at 32-bit by casting to u64

 arch/x86/kvm/vmx/nested.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
arch/x86/kvm/vmx/nested.c +1 -1
index f1a69117ac0f..1a51bff129a8 100644
@@ -4115,7 +4115,7 @@ int get_vmx_mem_address(struct kvm_vcpu *vcpu, unsigned long exit_qualification,
 		 */
 		if (!(s.base == 0 && s.limit == 0xffffffff &&
 		     ((s.type & 8) || !(s.type & 4))))
-			exn = exn || (off + sizeof(u64) > s.limit);
+			exn = exn || ((u64)off + sizeof(u64) - 1 > s.limit);
 	}
 	if (exn) {
 		kvm_queue_exception_e(vcpu,
Signature · 1 line
2.21.0
Reviewed-by on v5 1/3
#1 lore raw
Quoted 18 lines · Eugene Korenevsky

Intel SDM vol. 3, 5.3:
The processor causes a
general-protection exception (or, if the segment is SS, a stack-fault exception) any time an attempt is made to access the following addresses in a segment:
- A byte at an offset greater than the effective limit
- A word at an offset greater than the (effective-limit – 1)
- A doubleword at an offset greater than the (effective-limit – 3)
- A quadword at an offset greater than the (effective-limit – 7)

Therefore, the generic limit checking error condition must be

exn = (off > limit + 1 - access_len) = (off + access_len - 1 > limit)

but not

exn = (off + access_len > limit)

Quoted 8 lines

as for now.

Also avoid integer overflow of off at 32-bit KVM by casting it to u64.

Note: access length is currently sizeof(u64) which is incorrect. This will be fixed in the subsequent patch.

When sending a new revision of a patch, please add any supplied tags to the changelog, e.g. Reviewed-by, Tested-by, Acked-by, etc... For example, my Reviewed-by for v4 of this patch.

The one situation where you don't want to carry tags is if you make non-trivial changes to a patch, e.g. person X reviews a patch but then it gets reworked based on feedback from person Y, in which case the Reviewed-by from X should be dropped as they haven't reviewed the new code.

Thans for the patches!