Re: [PATCH 07/16] cpu: Introduce CPUSystemOperations structure

[Claudio Fontana]

On 2/26/21 5:32 PM, Philippe Mathieu-Daudé wrote:
> Introduce a structure to hold handler specific to sysemu.
> 
> Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
> ---
>  include/hw/core/cpu.h | 9 +++++++++
>  1 file changed, 9 insertions(+)
> 
> diff --git a/include/hw/core/cpu.h b/include/hw/core/cpu.h
> index b12028c3c03..ab89235cb45 100644
> --- a/include/hw/core/cpu.h
> +++ b/include/hw/core/cpu.h
> @@ -80,6 +80,12 @@ struct TCGCPUOps;
>  /* see accel-cpu.h */
>  struct AccelCPUClass;
>  
> +/*
> + * struct CPUSystemOperations: System operations specific to a CPU class
> + */
> +typedef struct CPUSystemOperations {

I'd suggest for consistency and brevity:

SysemuCPUOps .

Let me tell you my full train of thought, just so you understand exactly where 
I am coming from:

in my view all this camelcase is nonsense, as is typedef. They just cause 
problems and solve none.

However, considering the existing QEMU conventions, the existence of QOM,
for consistency with the rest of the QEMU code base, especially when looking at 
stuff around CPUClass,
the convention is to use this camel case stuff for objects, so in lieu of

tcg_cpu_ops

I went with TCGCPUOps

(TCG is the standard way to call tcg when case is an option, same for CPU, then 
Ops).

Here for consistency I would say:

Sysemu (as we are standardizing on calling system emulation/virtualization 
"sysemu")
+CPUOps as before.

=

SysemuCPUOps.

What do you think?

Ciao,

Claudio

> +} CPUSystemOperations;
> +
>  /**
>   * CPUClass:
>   * @class_by_name: Callback to map -cpu command line model name to an
> @@ -190,6 +196,9 @@ struct CPUClass {
>      bool gdb_stop_before_watchpoint;
>      struct AccelCPUClass *accel_cpu;
>  
> +    /* when system emulation is not available, this pointer is NULL */
> +    struct CPUSystemOperations system_ops;
> +
>      /* when TCG is not available, this pointer is NULL */
>      struct TCGCPUOps *tcg_ops;
>  };
>