Re: [kvm-unit-tests PATCH v8 01/12] s390x: Use PSW bits definitions in cstart

[Pierre Morel <pmorel@linux.ibm.com>]

On 2020-06-08 16:52, Thomas Huth wrote:
> On 08/06/2020 16.33, Pierre Morel wrote:
>>
>>
>> On 2020-06-08 10:43, Thomas Huth wrote:
>>> On 08/06/2020 10.12, Pierre Morel wrote:
>>>> This patch defines the PSW bits EA/BA used to initialize the PSW masks
>>>> for exceptions.
>>>>
>>>> Since some PSW mask definitions exist already in arch_def.h we add these
>>>> definitions there.
>>>> We move all PSW definitions together and protect assembler code against
>>>> C syntax.
>>>>
>>>> Signed-off-by: Pierre Morel <pmorel@linux.ibm.com>
>>>> Reviewed-by: Janosch Frank <frankja@linux.ibm.com>
>>>> ---
>>>>    lib/s390x/asm/arch_def.h | 15 +++++++++++----
>>>>    s390x/cstart64.S         | 15 ++++++++-------
>>>>    2 files changed, 19 insertions(+), 11 deletions(-)
>>>>
>>>> diff --git a/lib/s390x/asm/arch_def.h b/lib/s390x/asm/arch_def.h
>>>> index 1b3bb0c..5388114 100644
>>>> --- a/lib/s390x/asm/arch_def.h
>>>> +++ b/lib/s390x/asm/arch_def.h
>>>> @@ -10,15 +10,21 @@
>>>>    #ifndef _ASM_S390X_ARCH_DEF_H_
>>>>    #define _ASM_S390X_ARCH_DEF_H_
>>>>    +#define PSW_MASK_EXT            0x0100000000000000UL
>>>> +#define PSW_MASK_DAT            0x0400000000000000UL
>>>> +#define PSW_MASK_SHORT_PSW        0x0008000000000000UL
>>>> +#define PSW_MASK_PSTATE            0x0001000000000000UL
>>>> +#define PSW_MASK_BA            0x0000000080000000UL
>>>> +#define PSW_MASK_EA            0x0000000100000000UL
>>>> +
>>>> +#define PSW_EXCEPTION_MASK    (PSW_MASK_EA | PSW_MASK_BA)
>>>
>>> PSW_EXCEPTION_MASK sounds a little bit unfortunate - that term rather
>>> reminds me of something that disables some interrupts
>>> ... in case you
>>> respin, maybe rather use something like "PSW_EXC_ADDR_MODE" ?
>>
>> EXCEPTIONS_PSW_MASK ?
> 
> I think it is the _MASK suffix that mainly bugs me here, since this is
> not a define that you normally use for extracting the bits from a PSW...
> so EXCEPTIONS_PSW without _MASK would be fine for me... but as long as
> I'm the only one who has a strange feeling about this, it's also ok if
> you keep the current name.
> 
>   Thomas
> 

The _MASK is because it is applied to the psw.mask and not to the 
psw.addr part.

But I agree that the name is not good, to keep the naming convention, 
may be it should be:

PSW_MASK_ON_EXCEPTION

beginning with PSW_MASK_ like all other psw.mask definitions and 
ON_EXCEPTION clearly define when it is used.

Pierre

-- 
Pierre Morel
IBM Lab Boeblingen