Re: [Qemu-ppc] [PATCH 1/4] hw/ppc/spapr: Add h_set_sprg0 hypercall

[Thomas Huth]

On 11.02.2016 00:30, David Gibson wrote:
> On Wed, Feb 10, 2016 at 07:09:09PM +0100, Thomas Huth wrote:
>> This is a very simple hypercall that only sets up the SPRG0
>> register for the guest (since writing to SPRG0 was only permitted
>> to the hypervisor in older versions of the PowerISA).
>>
>> Signed-off-by: Thomas Huth <address@hidden>
>> ---
>>  hw/ppc/spapr_hcall.c | 15 +++++++++++++--
>>  1 file changed, 13 insertions(+), 2 deletions(-)
>>
>> diff --git a/hw/ppc/spapr_hcall.c b/hw/ppc/spapr_hcall.c
>> index 12f8c33..58103ef 100644
>> --- a/hw/ppc/spapr_hcall.c
>> +++ b/hw/ppc/spapr_hcall.c
>> @@ -332,6 +332,15 @@ static target_ulong h_read(PowerPCCPU *cpu, 
>> sPAPRMachineState *spapr,
>>      return H_SUCCESS;
>>  }
>>  
>> +static target_ulong h_set_sprg0(PowerPCCPU *cpu, sPAPRMachineState *spapr,
>> +                                target_ulong opcode, target_ulong *args)
>> +{
>> +    CPUState *cs = CPU(cpu);
>> +
>> +    set_spr(cs, SPR_SPRG0, args[0], -1L);
> 
> This looks correct, but I think set_spr() is serious overkill here.
> It does some fancy synchronization designed for setting one cpu's SPR
> from an hcall executed on a different CPU.  In this case the calling
> CPU is just setting its own SPRG0, so just
>       cpu_synchronize_state()
>       env->spr[SPR_SPRG0] = XXX
> 
> Should be sufficient.

AFAIK the synchronization stuff is skipped when set_spr() runs already
on the destination CPU, but ok, since h-calls should be fast, I can
change this anyway to save some precious cycles.

 Thomas

signature.asc
Description: OpenPGP digital signature