On 11/4/22 21:58, Peter Maydell wrote:
OK. Do we ever do anything with the attrs for a phys tlb lookup
except use them internally for details of the stage 2 tlb walk ?
I guess they get used for the memory transaction to do the walk.
That matches the old code that just had a local MemTxAttrs in
On Thu, 3 Nov 2022 at 20:30, Richard Henderson
wrote:
>
> On 11/4/22 00:19, Peter Maydell wrote:
> >> @@ -2671,6 +2671,13 @@ static bool get_phys_addr_with_struct(CPUARMState
> >> *env, S1Translate *ptw,
> >> bool is_secure = ptw->in_secure;
> >> ARMMMUIdx s1_mmu_idx;
> >>
> >> +
On 11/4/22 00:19, Peter Maydell wrote:
@@ -2671,6 +2671,13 @@ static bool get_phys_addr_with_struct(CPUARMState *env,
S1Translate *ptw,
bool is_secure = ptw->in_secure;
ARMMMUIdx s1_mmu_idx;
+/*
+ * The page table entries may downgrade secure to non-secure, but
+ *
On Thu, 3 Nov 2022 at 13:19, Peter Maydell wrote:
>
> On Wed, 2 Nov 2022 at 05:47, Richard Henderson
> wrote:
> >
> > Reversed the sense of non-secure in get_phys_addr_lpae,
> > and failed to initialize attrs.secure for ARMMMUIdx_Phys_S.
> >
> > Fixes: 48da29e4 ("target/arm: Add ptw_idx to
On Wed, 2 Nov 2022 at 05:47, Richard Henderson
wrote:
>
> Reversed the sense of non-secure in get_phys_addr_lpae,
> and failed to initialize attrs.secure for ARMMMUIdx_Phys_S.
>
> Fixes: 48da29e4 ("target/arm: Add ptw_idx to S1Translate")
> Resolves:
On 2/11/22 06:47, Richard Henderson wrote:
Reversed the sense of non-secure in get_phys_addr_lpae,
and failed to initialize attrs.secure for ARMMMUIdx_Phys_S.
Fixes: 48da29e4 ("target/arm: Add ptw_idx to S1Translate")
Resolves: https://gitlab.com/qemu-project/qemu/-/issues/1293
Thanks!
Reversed the sense of non-secure in get_phys_addr_lpae,
and failed to initialize attrs.secure for ARMMMUIdx_Phys_S.
Fixes: 48da29e4 ("target/arm: Add ptw_idx to S1Translate")
Resolves: https://gitlab.com/qemu-project/qemu/-/issues/1293
Signed-off-by: Richard Henderson
---
target/arm/ptw.c | 15