[PATCH] arm/arm64: KVM: Feed initialized memory to MMIO accesses

Subsystems: arm port, the rest

STALE3811d

3 messages, 2 authors, 2016-02-24 · open the first message on its own page

[PATCH] arm/arm64: KVM: Feed initialized memory to MMIO accesses

From: Marc Zyngier <hidden>
Date: 2016-02-17 15:34:20

On an MMIO access, we always copy the on-stack buffer info
the shared "run" structure, even if this is a read access.
This ends up leaking up to 8 bytes of uninitialized memory
into userspace.

An obvious fix for this one is to only perform the copy if
this is an actual write.

Signed-off-by: Marc Zyngier <redacted>
---
 arch/arm/kvm/mmio.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/arch/arm/kvm/mmio.c b/arch/arm/kvm/mmio.c
index 7f33b20..0f6600f 100644
--- a/arch/arm/kvm/mmio.c
+++ b/arch/arm/kvm/mmio.c
@@ -206,7 +206,8 @@ int io_mem_abort(struct kvm_vcpu *vcpu, struct kvm_run *run,
 	run->mmio.is_write	= is_write;
 	run->mmio.phys_addr	= fault_ipa;
 	run->mmio.len		= len;
-	memcpy(run->mmio.data, data_buf, len);
+	if (is_write)
+		memcpy(run->mmio.data, data_buf, len);
 
 	if (!ret) {
 		/* We handled the access successfully in the kernel. */
-- 
2.1.4

[PATCH] arm/arm64: KVM: Feed initialized memory to MMIO accesses

From: Christoffer Dall <hidden>
Date: 2016-02-24 11:40:44

On Wed, Feb 17, 2016 at 03:34:20PM +0000, Marc Zyngier wrote:
On an MMIO access, we always copy the on-stack buffer info
the shared "run" structure, even if this is a read access.
This ends up leaking up to 8 bytes of uninitialized memory
into userspace.
I think it only leaks 'len' bytes to userspace ;)
An obvious fix for this one is to only perform the copy if
this is an actual write.
Reviewed-by: Christoffer Dall <redacted>
quoted hunk
Signed-off-by: Marc Zyngier <redacted>
---
 arch/arm/kvm/mmio.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/arch/arm/kvm/mmio.c b/arch/arm/kvm/mmio.c
index 7f33b20..0f6600f 100644
--- a/arch/arm/kvm/mmio.c
+++ b/arch/arm/kvm/mmio.c
@@ -206,7 +206,8 @@ int io_mem_abort(struct kvm_vcpu *vcpu, struct kvm_run *run,
 	run->mmio.is_write	= is_write;
 	run->mmio.phys_addr	= fault_ipa;
 	run->mmio.len		= len;
-	memcpy(run->mmio.data, data_buf, len);
+	if (is_write)
+		memcpy(run->mmio.data, data_buf, len);
 
 	if (!ret) {
 		/* We handled the access successfully in the kernel. */
-- 
2.1.4

[PATCH] arm/arm64: KVM: Feed initialized memory to MMIO accesses

From: Marc Zyngier <hidden>
Date: 2016-02-24 12:06:52

On 24/02/16 11:40, Christoffer Dall wrote:
On Wed, Feb 17, 2016 at 03:34:20PM +0000, Marc Zyngier wrote:
quoted
On an MMIO access, we always copy the on-stack buffer info
the shared "run" structure, even if this is a read access.
This ends up leaking up to 8 bytes of uninitialized memory
into userspace.
I think it only leaks 'len' bytes to userspace ;)
quoted
An obvious fix for this one is to only perform the copy if
this is an actual write.
Reviewed-by: Christoffer Dall <redacted>
Thanks. I've pushed this onto master, with a view of sending a PR to
Paolo this evening (hopefully the last one for this cycle).

	M.
-- 
Jazz is not dead. It just smells funny...
Keyboard shortcuts
hback out one level
jnext message in thread
kprevious message in thread
ldrill in
Escclose help / fold thread tree
?toggle this help