Re: [RFC v5 09/12] module: introduce MODULE_INIT_ACCEL_CPU

[Claudio Fontana]

On 11/25/20 10:30 AM, Paolo Bonzini wrote:
> On 25/11/20 10:21, Claudio Fontana wrote:
>> Hi Paolo,
>>
>> in RFC v5 , module init for ACCEL_CPU is not conditional anymore, right?
>> But the fact that its behavior depends on current_accel() still disqualifies 
>> it?
>> It is called right after the accelerator is chosen and initialized
>> in RFC v5, this still is "in the middle of the machine creation sequence"?
> Yes, machine creation basically starts after command line parsing, or 
> perhaps even _with_ command line parsing.  Basically once the user can 
> control the flow it is already too late.
> 
>> I am trying to find the actual things to fix, since when doing RFC
>> v5  I tried to specifically address two points:
>>
>> 1) no if () inside module init functions
>>
>> 2) no proliferation of module init functions
>>
>> which I accomplished via AccelClass extension to user mode, current_accel(), 
>> and class lookup.
> 
> Yes, the rest is great, I'm just not sure that MODULE_INIT_ACCEL_CPU is 
> useful and if virtual functions on accel and CPU_RESOLVING_TYPE can 
> achieve the same.
> 
>> If MODULE_INIT_ACCEL_CPU remains an option, where would you like to see the 
>> call so that it is not "in the middle"?
> 
> No later than the runstate_init() call, roughly.
> 
> Paolo
> 

Hi Paolo, not super-related to the context above,

during the implementation I am trying to offer

accel/accel-softmmu.c
accel/accel-common.c
accel/accel-user.c

But I don't seem able to use CONFIG_USER_ONLY in accel-common.c .

in accel/meson.build I am saying

 common_ss.add(files('accel-common.c'))
 softmmu_ss.add(files('accel-softmmu.c'))
 user_ss.add(files('accel-user.c'))

But this doesn't work, if I use common_ss. If I use specific_ss, it works.

So the term "common" is a bit overloaded, it means stuff for libcommon and 
such, not common between softmmu and user, right?

Ciao,

Claudio