On 5/20/23 02:45, Philippe Mathieu-Daudé wrote:
When multiple QOM types are registered in the same file, it is simpler to use the the DEFINE_TYPES() macro. Replace the type_init() / type_register_static() combination. This is in preparation of adding the OpenTitan machine type to this array in a pair of commits.
I wonder whether we should forbid type_init()/type_register_static() and force everyone to use DEFINE_TYPES(). We need less options when dealing with QOM. Reviewed-by: Daniel Henrique Barboza <dbarb...@ventanamicro.com>
Signed-off-by: Philippe Mathieu-Daudé <phi...@linaro.org> --- hw/riscv/opentitan.c | 21 +++++++++------------ 1 file changed, 9 insertions(+), 12 deletions(-) diff --git a/hw/riscv/opentitan.c b/hw/riscv/opentitan.c index 2d21ee39c5..294955eeea 100644 --- a/hw/riscv/opentitan.c +++ b/hw/riscv/opentitan.c @@ -320,17 +320,14 @@ static void lowrisc_ibex_soc_class_init(ObjectClass *oc, void *data) dc->user_creatable = false; }-static const TypeInfo lowrisc_ibex_soc_type_info = {- .name = TYPE_RISCV_IBEX_SOC, - .parent = TYPE_DEVICE, - .instance_size = sizeof(LowRISCIbexSoCState), - .instance_init = lowrisc_ibex_soc_init, - .class_init = lowrisc_ibex_soc_class_init, +static const TypeInfo open_titan_types[] = { + { + .name = TYPE_RISCV_IBEX_SOC, + .parent = TYPE_DEVICE, + .instance_size = sizeof(LowRISCIbexSoCState), + .instance_init = lowrisc_ibex_soc_init, + .class_init = lowrisc_ibex_soc_class_init, + } };-static void lowrisc_ibex_soc_register_types(void)-{ - type_register_static(&lowrisc_ibex_soc_type_info); -} - -type_init(lowrisc_ibex_soc_register_types) +DEFINE_TYPES(open_titan_types)
