Re: [Qemu-devel] [PATCH 04/67] s390x: drop an unused include

[Philippe Mathieu-Daudé]

On 05/08/2018 10:50 AM, Thomas Huth wrote:
> On 08.05.2018 15:45, Cornelia Huck wrote:
>> On Tue, 8 May 2018 15:38:03 +0200
>> Thomas Huth <address@hidden> wrote:
>>
>>> On 08.05.2018 15:23, Cornelia Huck wrote:
>>>> On Fri, 4 May 2018 02:24:12 +0200
>>>> Thomas Huth <address@hidden> wrote:
>>>>   
>>>>> On 03.05.2018 21:50, Michael S. Tsirkin wrote:  
>>>>>> we just need a struct name, let's add a forward
>>>>>> declaration instead of an include.
>>>>>>
>>>>>> Signed-off-by: Michael S. Tsirkin <address@hidden>
>>>>>> ---
>>>>>>  include/hw/s390x/sclp.h | 3 ++-
>>>>>>  1 file changed, 2 insertions(+), 1 deletion(-)
>>>>>>
>>>>>> diff --git a/include/hw/s390x/sclp.h b/include/hw/s390x/sclp.h
>>>>>> index f9db243..6e65150 100644
>>>>>> --- a/include/hw/s390x/sclp.h
>>>>>> +++ b/include/hw/s390x/sclp.h
>>>>>> @@ -16,7 +16,8 @@
>>>>>>  
>>>>>>  #include "hw/sysbus.h"
>>>>>>  #include "hw/qdev.h"
>>>>>> -#include "target/s390x/cpu-qom.h"
>>>>>> +
>>>>>> +typedef struct CPUS390XState CPUS390XState;    
>>>>>
>>>>> IIRC you have to use include/qemu/typedefs.h instead to avoid trouble
>>>>> with older versions of GCC.  
>>>>
>>>> Hm, I'm wondering why we do the typedef in cpu-qom.h, while other
>>>> architectures do it in their cpu.h.  
>>>
>>> See:
>>>
>>> commit ef2974cc270d51959ce90df6b4d4d41635d7a603
>>> Author: David Hildenbrand <address@hidden>
>>> Date:   Wed Sep 13 15:24:02 2017 +0200
>>>
>>>     target/s390x: move some s390x typedefs to cpu-qom.h
>>>     
>>>     This allows us to drop inclusion of cpu_models.h in cpu-qom.h, and
>>>     prepares for using cpu-qom.h as a s390 specific version of typedefs.h
>>>     
>>>     Signed-off-by: David Hildenbrand <address@hidden>
>>>     Message-Id: <address@hidden>
>>>     Reviewed-by: Thomas Huth <address@hidden>
>>>     Signed-off-by: Cornelia Huck <address@hidden>
>>>
>>>  Thomas
>>
>> Gargh, this is all very confusing...
> 
> If you'd ask me, I'd say we should get rid of the typedefs and do it the
> Linux kernel way and enforce using "struct xyz" everywhere, then you
> also do not have this problem with typedefs.h anymore ... but well, so
> far it seems as I'm still part of a minority with this opinion here.

Maybe not getting rid of the typedefs, but I agree with removing typedefs.h.