Re: [PATCH 17/22] target/arm: Use get_phys_addr_with_struct for stage2

[Peter Maydell]

On Mon, 20 Feb 2023 at 22:15, Richard Henderson
<richard.henderson@linaro.org> wrote:
>
> On 2/10/23 03:28, Peter Maydell wrote:
> > On Tue, 24 Jan 2023 at 00:01, Richard Henderson
> > <richard.henderson@linaro.org> wrote:
> >>
> >> This fixes a bug in which we failed to initialize
> >> the result attributes properly after the memset.
> >>
> >> Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
> >> ---
> >>   target/arm/ptw.c | 13 +------------
> >>   1 file changed, 1 insertion(+), 12 deletions(-)
> >>
> >> diff --git a/target/arm/ptw.c b/target/arm/ptw.c
> >> index eaa47f6b62..3205339957 100644
> >> --- a/target/arm/ptw.c
> >> +++ b/target/arm/ptw.c
> >> @@ -32,12 +32,6 @@ typedef struct S1Translate {
> >>       void *out_host;
> >>   } S1Translate;
> >>
> >> -static bool get_phys_addr_lpae(CPUARMState *env, S1Translate *ptw,
> >> -                               uint64_t address,
> >> -                               MMUAccessType access_type,
> >> -                               GetPhysAddrResult *result, ARMMMUFaultInfo 
> >> *fi)
> >> -    __attribute__((nonnull));
> >
> > The definition of the function doesn't have the __attribute__,
> > so if we drop this forward declaration we need to move the attribute.
>
> Eh.  It was useful as an intermediary during one of the ptw reorgs, but now 
> that we've
> eliminated the use case in which NULL had been passed, it can go away.  I 
> assume you'd
> prefer that as a separate patch?

Yes, if we want to deliberately drop the attribute we should
do that separately with justification for why it's not needed.

thanks
-- PMM