Re: [Qemu-devel] [PATCH 02/12] qtest: add spapr hypercall support

[Alexander Graf]

On 20.06.2013, at 20:58, Anthony Liguori wrote:

> Alexander Graf <address@hidden> writes:
> 
>> Am 20.06.2013 um 17:42 schrieb Anthony Liguori <address@hidden>:
>> 
>>> Andreas Färber <address@hidden> writes:
>>> 
>>>> Am 19.06.2013 22:40, schrieb Anthony Liguori:
>>>>> Signed-off-by: Anthony Liguori <address@hidden>
>>>>> ---
>>>>> qtest.c          | 29 +++++++++++++++++++++++++++++
>>>>> tests/libqtest.c | 18 ++++++++++++++++++
>>>>> tests/libqtest.h | 46 ++++++++++++++++++++++++++++++++++++++++++++++
>>>>> 3 files changed, 93 insertions(+)
>>>>> 
>>>>> diff --git a/qtest.c b/qtest.c
>>>>> index 07a9612..f8c8f44 100644
>>>>> --- a/qtest.c
>>>>> +++ b/qtest.c
>>>>> @@ -19,6 +19,9 @@
>>>>> #include "hw/irq.h"
>>>>> #include "sysemu/sysemu.h"
>>>>> #include "sysemu/cpus.h"
>>>>> +#ifdef TARGET_PPC64
>>>>> +#include "hw/ppc/spapr.h"
>>>>> +#endif
>>>>> 
>>>>> #define MAX_IRQ 256
>>>>> 
>>>>> @@ -141,6 +144,13 @@ static bool qtest_opened;
>>>>> * where NUM is an IRQ number.  For the PC, interrupts can be intercepted
>>>>> * simply with "irq_intercept_in ioapic" (note that IRQ0 comes out with
>>>>> * NUM=0 even though it is remapped to GSI 2).
>>>>> + *
>>>>> + * Platform specific (sPAPR):
>>>>> + *
>>>>> + *  > papr_hypercall NR ARG0 ARG1 ... ARG8
>>>> 
>>>> The functions are called spapr_hcall*() but the protocol uses
>>>> papr_hypercall?
>>> 
>>> The discrepancy is inherited in the KVM vs. QEMU interfaces.  It's
>>> called papr_hypercall in the KVM interface vs. spapr in QEMU.
>>> 
>>> I honestly don't know what the distinction between spapr and papr is.
>> 
>> PAPR is what PAPR calls itself. However, there is also an ePAPR for
>> BookE, so in order to distinguish the 2 more easily, we named the
>> server version spapr wherever we remembered to.
> 
> So does it make sense to have papr_hypercall()?  Do hypercalls exist
> with the virtualization extensions on BookE?

papr_hypercall() really means spapr_hypercall() :). I don't think we should 
mangle ePAPR and sPAPR together.


Alex