Re: [PATCH] target-info: dispatch CPU QMP, dump, monitor, and Angel semihosting

["Daniel P. Berrangé" <berrange@redhat.com>]

On Thu, Oct 15, 2026 at 12:41:02PM +0800, Yonggang Luo wrote:
> Put query-cpu-definitions, dump notes, monitor, and Angel semihosting on
> TargetOps[SYS_EMU_TARGET__MAX] and look up by target_arch().
> Keep the table in target-ops.c and with a stub
> 
> - Register with a QEMU_ARCH_* bitmask so shared base-arch objects
>   can fill both slots (e.g. QEMU_ARCH_ARM is ARM|AARCH64).
> - TARGET_INFO_CPU_OP is type_init. A NULL member means no handler.
> - Drop stubs/qmp-cpu.c
>        stubs/hmp-cmd-info_tlb.c
>        stubs/hmp-cmd-info_mem.c
>        stubs/dump.c
> 
> Signed-off-by: Yonggang Luo <luoyonggang@gmail.com>
> +typedef struct TargetOps {
> +    void (*info_tlb)(MonitorHMP *hmp, const QDict *qdict);
> +    void (*info_mem)(MonitorHMP *hmp, const QDict *qdict);
> +    CpuDefinitionInfoList *(*query_cpu_definitions)(Error **errp);
> +    CpuModelExpansionInfo *(*query_cpu_model_expansion)(
> +        CpuModelExpansionType type, CpuModelInfo *model, Error **errp);
> +    int (*get_dump_info)(ArchDumpInfo *info,
> +                         const struct GuestPhysBlockList *guest_phys_blocks);
> +    ssize_t (*get_note_size)(int class, int machine, int nr_cpus);
> +    const CPUSemihostingOps *semihosting;
> +} TargetOps;

IMHO, TargetOps shouldn't exist and this kind of thing should
be done as virtual methods aginst the CPUClass, or failing
that, against the MachineClass.


With regards,
Daniel
-- 
|: https://berrange.com       ~~        https://hachyderm.io/@berrange :|
|: https://libvirt.org          ~~          https://entangle-photo.org :|
|: https://pixelfed.art/berrange   ~~    https://fstop138.berrange.com :|