Re: [PATCH v3 02/10] include: Move compat_timespec/ timeval to compat_time.h

[Christian Borntraeger]

On 03/06/2018 01:46 PM, Arnd Bergmann wrote:
> On Mon, Mar 5, 2018 at 10:30 AM, Christian Borntraeger
> <borntraeger@xxxxxxxxxx> wrote:
>> On 01/16/2018 03:18 AM, Deepa Dinamani wrote:
>>> All the current architecture specific defines for these
>>> are the same. Refactor these common defines to a common
>>> header file.
>>>
>>> The new common linux/compat_time.h is also useful as it
>>> will eventually be used to hold all the defines that
>>> are needed for compat time types that support non y2038
>>> safe types. New architectures need not have to define these
>>> new types as they will only use new y2038 safe syscalls.
>>> This file can be deleted after y2038 when we stop supporting
>>> non y2038 safe syscalls.
>>
>> You are now include a <linux/*.h> from several asm files
>> (
>>  arch/arm64/include/asm/stat.h
>>  arch/s390/include/asm/elf.h
>>  arch/x86/include/asm/ftrace.h
>>  arch/x86/include/asm/sys_ia32.h
>> )
>> It works, and it is done in many places, but it looks somewhat weird.
>> Would it make sense to have an asm-generic/compate-time.h instead? Asking for
>> opinions here.
> 
> I don't think we have such a rule. If a header file is common to all
> architectures (i.e. no architecture uses a different implementation),
> it should be in include/linux rather than include/asm-generic, regardless
> of whether it can be used by assembler files or not.
> 
>>> --- a/drivers/s390/net/qeth_core_main.c
>>> +++ b/drivers/s390/net/qeth_core_main.c
>>> @@ -32,7 +32,7 @@
>>>  #include <asm/chpid.h>
>>>  #include <asm/io.h>
>>>  #include <asm/sysinfo.h>
>>> -#include <asm/compat.h>
>>> +#include <linux/compat.h>
>>>  #include <asm/diag.h>
>>>  #include <asm/cio.h>
>>>  #include <asm/ccwdev.h>
>>
>> Can you move that into the other includes (where all the other <linux/*> includes are.
> 
> Good catch, this is definitely a rule we have ;-)

FWIW, this was also broken for 
arch/x86/include/asm/sys_ia32.h