kvm_vncr_tlb_lookup() is supposed to return true when the cached VNCR
TLB entry is valid for the current context. For non-Global entries, that
means the entry’s ASID must match the current ASID.

The current code returns true when the ASIDs do *not* match, which
inverts the logic. This is a potential vulnerability:

- Valid entries are ignored and we fall back to kvm_translate_vncr(),
  hurting performance.
- Mismatched entries are treated as permission faults (-EPERM) instead
  of triggering a fresh translation.
- This can also cause stale translations to be (wrongly) considered
  valid across address spaces.

Flip the predicate so non-Global entries only hit when ASIDs match.

Reported-by: Team 0xB6 in bob14
  DongHa Lee (@GAP-dev)
  Gyujeong Jin (@gyutrange)
  Daehyeon Ko (@4ncienth)
  Geonha Lee (@leegn4a)
  Hyungyu Oh (@DQPC_lover)
  Jaewon Yang (@R4mbb)

Signed-off-by: Geonha Lee <w1nsom3gna@korea.ac.kr>
---
 arch/arm64/kvm/nested.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/arch/arm64/kvm/nested.c b/arch/arm64/kvm/nested.c
index 77db81bae86f..24eab94d7d7f 100644
--- a/arch/arm64/kvm/nested.c
+++ b/arch/arm64/kvm/nested.c
@@ -1276,7 +1276,7 @@ static bool kvm_vncr_tlb_lookup(struct kvm_vcpu *vcpu)
 		    !(tcr & TCR_ASID16))
 			asid &= GENMASK(7, 0);
 
-		return asid != vt->wr.asid;
+		return asid == vt->wr.asid;
 	}
 
 	return true;
-- 
2.43.0

On Thu, 04 Sep 2025 00:04:21 +0900, Geonha Lee wrote:
> kvm_vncr_tlb_lookup() is supposed to return true when the cached VNCR
> TLB entry is valid for the current context. For non-Global entries, that
> means the entry’s ASID must match the current ASID.
> 
> The current code returns true when the ASIDs do *not* match, which
> inverts the logic. This is a potential vulnerability:
> 
> [...]

Applied to fixes, thanks!

[1/1] KVM: arm64: nested: fix VNCR TLB ASID match logic for non-Global entries
      https://git.kernel.org/kvmarm/kvmarm/c/06f66db9bda4

--
Best,
Oliver

On Wed, 03 Sep 2025 16:04:21 +0100,
Geonha Lee <w1nsom3gna@korea.ac.kr> wrote:
> 
> kvm_vncr_tlb_lookup() is supposed to return true when the cached VNCR
> TLB entry is valid for the current context. For non-Global entries, that
> means the entry’s ASID must match the current ASID.
> 
> The current code returns true when the ASIDs do *not* match, which
> inverts the logic. This is a potential vulnerability:
> 
> - Valid entries are ignored and we fall back to kvm_translate_vncr(),
>   hurting performance.
> - Mismatched entries are treated as permission faults (-EPERM) instead
>   of triggering a fresh translation.
> - This can also cause stale translations to be (wrongly) considered
>   valid across address spaces.

I don't immediately see the vulnerability on the host. In the guest,
yes, absolutely.

> 
> Flip the predicate so non-Global entries only hit when ASIDs match.
> 
> Reported-by: Team 0xB6 in bob14
>   DongHa Lee (@GAP-dev)
>   Gyujeong Jin (@gyutrange)
>   Daehyeon Ko (@4ncienth)
>   Geonha Lee (@leegn4a)
>   Hyungyu Oh (@DQPC_lover)
>   Jaewon Yang (@R4mbb)

Reported-by: has a specific meaning, and needs addresses. Oliver, can
you change this to some sort of attribution?

> 
> Signed-off-by: Geonha Lee <w1nsom3gna@korea.ac.kr>
> ---
>  arch/arm64/kvm/nested.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/arch/arm64/kvm/nested.c b/arch/arm64/kvm/nested.c
> index 77db81bae86f..24eab94d7d7f 100644
> --- a/arch/arm64/kvm/nested.c
> +++ b/arch/arm64/kvm/nested.c
> @@ -1276,7 +1276,7 @@ static bool kvm_vncr_tlb_lookup(struct kvm_vcpu *vcpu)
>  		    !(tcr & TCR_ASID16))
>  			asid &= GENMASK(7, 0);
>  
> -		return asid != vt->wr.asid;
> +		return asid == vt->wr.asid;
>  	}
>  
>  	return true;

Yup, looks correct to me. Thanks again for fixing it.

Reviewed-by: Marc Zyngier <maz@kernel.org>

	M.

-- 
Jazz isn't dead. It just smells funny.