On 8/30/19 2:07 PM, David Hildenbrand wrote:
> On 26.08.19 18:35, Janosch Frank wrote:
>> For now let's concentrate on the error conditions.
>>
>> Signed-off-by: Janosch Frank <frankja@xxxxxxxxxxxxx>
>> ---
[...]
>> +static inline unsigned long stsi_get_fc(void *addr)
>> +{
>> + register unsigned long r0 asm("0") = 0;
>> + register unsigned long r1 asm("1") = 0;
>> + int cc;
>> +
>> + asm volatile("stsi 0(%3)\n"
>> + "ipm %[cc]\n"
>> + "srl %[cc],28\n"
>> + : "+d" (r0), [cc] "=d" (cc)
>> + : "d" (r1), "a" (addr)
>
> maybe [addr], so you can avoid the %3 above
Sure, maybe Thomas can also fix that on picking for the previous patch?
>
>> + : "cc", "memory");
>> + assert(!cc);
>> + return r0 >> 28;
>
> I think I'd prefer "get_configuration_level()" and move it to an header
> - because the fc actually allows more values (0, 15 ...) - however the
> level can be used as an fc.
The rename works for me, but that's currently used only once, so why
should it go to a header file?
I though about starting lib/s390x/asm/misc-instr.h if we have enough (>=
2) instruction definitions which are shared.
>
>
>> +}
>> +
>> +static void test_fc(void)
>> +{
>> + report("invalid fc", stsi(pagebuf, 7, 0, 0) == 3);
>> + report("query fc >= 2", stsi_get_fc(pagebuf) >= 2);
>> +}
>> +
>> +int main(void)
>> +{
>> + report_prefix_push("stsi");
>> + test_priv();
>> + test_specs();
>> + test_fc();
>> + return report_summary();
>> +}
>> diff --git a/s390x/unittests.cfg b/s390x/unittests.cfg
>> index 9dd288a..cc79a4e 100644
>> --- a/s390x/unittests.cfg
>> +++ b/s390x/unittests.cfg
>> @@ -68,3 +68,6 @@ file = cpumodel.elf
>> [diag288]
>> file = diag288.elf
>> extra_params=-device diag288,id=watchdog0 --watchdog-action inject-nmi
>> +
>> +[stsi]
>> +file = stsi.elf
>>
>
> Apart from that
>
> Reviewed-by: David Hildenbrand <david@xxxxxxxxxx>
>
Attachment:
signature.asc
Description: OpenPGP digital signature
