On Fri, Apr 11, 2014 at 2:15 AM, Peter Maydell <peter.maydell@xxxxxxxxxx> wrote:
> The AArch64 implementation of the set_pc method needs to be updated to
> handle the possibility that the CPU is in AArch32 mode; otherwise there
> are weird crashes when doing interprocessing in system emulation mode
> when an interrupt occurs and we fail to resynchronize the 32-bit PC
> with the TB we need to execute next.
>
> Signed-off-by: Peter Maydell <peter.maydell@xxxxxxxxxx>
> Reviewed-by: Alex Bennée <alex.bennee@xxxxxxxxxx>
Reviewed-by: Peter Crosthwaite <peter.crosthwaite@xxxxxxxxxx>
> ---
> target-arm/cpu64.c | 12 ++++++++----
> 1 file changed, 8 insertions(+), 4 deletions(-)
>
> diff --git a/target-arm/cpu64.c b/target-arm/cpu64.c
> index 9a0c431..c673ac2 100644
> --- a/target-arm/cpu64.c
> +++ b/target-arm/cpu64.c
> @@ -172,11 +172,15 @@ static void aarch64_cpu_finalizefn(Object *obj)
> static void aarch64_cpu_set_pc(CPUState *cs, vaddr value)
> {
> ARMCPU *cpu = ARM_CPU(cs);
> - /*
> - * TODO: this will need updating for system emulation,
> - * when the core may be in AArch32 mode.
> + /* It's OK to look at env for the current mode here, because it's
> + * never possible for an AArch64 TB to chain to an AArch32 TB.
> + * (Otherwise we would need to use synchronize_from_tb instead.)
> */
> - cpu->env.pc = value;
> + if (is_a64(&cpu->env)) {
> + cpu->env.pc = value;
> + } else {
> + cpu->env.regs[15] = value;
> + }
> }
>
> static void aarch64_cpu_class_init(ObjectClass *oc, void *data)
> --
> 1.9.1
>
>
_______________________________________________
kvmarm mailing list
kvmarm@xxxxxxxxxxxxxxxxxxxxx
https://lists.cs.columbia.edu/mailman/listinfo/kvmarm
