On 9/18/23 09:02, Philippe Mathieu-Daudé wrote:
> Architecture specific hardware doesn't have a particular dependency
> on the accelerator vCPU (created with cpu_exec_realizefn), and can
> be initialized *after* the vCPU is realized. Doing so allows further
> generic API simplification (in few commits).
>
> Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
> ---
> target/hppa/cpu.c | 4 ++--
> 1 file changed, 2 insertions(+), 2 deletions(-)
>
> diff --git a/target/hppa/cpu.c b/target/hppa/cpu.c
> index 49082bd2ba..b0d106b6c7 100644
> --- a/target/hppa/cpu.c
> +++ b/target/hppa/cpu.c
> @@ -131,8 +131,6 @@ static void hppa_cpu_realizefn(DeviceState *dev, Error **errp)
> return;
> }
>
> - acc->parent_realize(dev, errp);
> -
> #ifndef CONFIG_USER_ONLY
> {
> HPPACPU *cpu = HPPA_CPU(cs);
> @@ -140,6 +138,8 @@ static void hppa_cpu_realizefn(DeviceState *dev, Error **errp)
> hppa_cpu_alarm_timer, cpu);
> }
> #endif
> +
> + acc->parent_realize(dev, errp);
> }
>
> static void hppa_cpu_initfn(Object *obj)
This appears to delay final realization of the vcpu, not advance it...
r~