On 11/28/19 3:40 PM, Thomas Huth wrote:
> On 20/11/2019 12.43, Janosch Frank wrote:
>> For protected guests the IPIB is written/read to/from the sattelite
> 
> satellite
> 
>> block, so we need to make those accesses virtual to make them go
>> through KBM mem ops.
> 
> What's KBM ?

A totally new hypervisor that's much faster than KVM because b comes
before v in the alphabet.

> 
>> Signed-off-by: Janosch Frank <frankja@linux.ibm.com>
>> ---
>>  target/s390x/diag.c | 26 +++++++++++++++++++++++---
>>  1 file changed, 23 insertions(+), 3 deletions(-)
>>
>> diff --git a/target/s390x/diag.c b/target/s390x/diag.c
>> index db6d79cef3..d96d8bdc6c 100644
>> --- a/target/s390x/diag.c
>> +++ b/target/s390x/diag.c
>> @@ -88,6 +88,7 @@ static int diag308_parm_check(CPUS390XState *env, uint64_t r1, uint64_t addr,
>>  void handle_diag_308(CPUS390XState *env, uint64_t r1, uint64_t r3, uintptr_t ra)
>>  {
>>      CPUState *cs = env_cpu(env);
>> +    S390CPU *cpu = S390_CPU(cs);
>>      uint64_t addr =  env->regs[r1];
>>      uint64_t subcode = env->regs[r3];
>>      IplParameterBlock *iplb;
>> @@ -118,14 +119,27 @@ void handle_diag_308(CPUS390XState *env, uint64_t r1, uint64_t r3, uintptr_t ra)
>>          if (diag308_parm_check(env, r1, addr, ra, false)) {
>>              return;
>>          }
>> +
>>          iplb = g_new0(IplParameterBlock, 1);
>> -        cpu_physical_memory_read(addr, iplb, sizeof(iplb->len));
>> +        if (!env->pv) {
>> +            cpu_physical_memory_read(addr, iplb, sizeof(iplb->len));
>> +        } else {
>> +            s390_cpu_virt_mem_read(cpu, 0, 0, iplb, sizeof(iplb->len));
>> +            s390_cpu_virt_mem_handle_exc(cpu, ra);
> 
> I'm looking forward to protected virt support in TCG ;-)

Who doesn't?

> 
>> +        }
>> +
>>          if (!iplb_valid_len(iplb)) {
>>              env->regs[r1 + 1] = DIAG_308_RC_INVALID;
>>              goto out;
>>          }
>>  
>> -        cpu_physical_memory_read(addr, iplb, be32_to_cpu(iplb->len));
>> +        if (!env->pv) {
>> +            cpu_physical_memory_read(addr, iplb, be32_to_cpu(iplb->len));
>> +        } else {
>> +            s390_cpu_virt_mem_read(cpu, 0, 0, iplb, be32_to_cpu(iplb->len));
>> +            s390_cpu_virt_mem_handle_exc(cpu, ra);
>> +        }
>> +
>>  
>>          if (!iplb_valid_ccw(iplb) && !iplb_valid_fcp(iplb) &&
>>              !(iplb_valid_se(iplb) && s390_ipl_pv_check_comp(iplb) >= 0)) {
>> @@ -149,7 +163,13 @@ out:
>>              iplb = s390_ipl_get_iplb();
>>          }
>>          if (iplb) {
>> -            cpu_physical_memory_write(addr, iplb, be32_to_cpu(iplb->len));
>> +            if (!env->pv) {
>> +                cpu_physical_memory_write(addr, iplb, be32_to_cpu(iplb->len));
>> +            } else {
>> +                s390_cpu_virt_mem_write(cpu, 0, 0, iplb,
>> +                                        be32_to_cpu(iplb->len));
>> +                s390_cpu_virt_mem_handle_exc(cpu, ra);
>> +            }
>>              env->regs[r1 + 1] = DIAG_308_RC_OK;
>>          } else {
>>              env->regs[r1 + 1] = DIAG_308_RC_NO_CONF;
>>
> 
> I wonder whether we maybe want to have some wrapper functions for these
> reads and writes, something like:

diag308 is currently the only instruction where we would need it.

sclp needs a new handler, because we can skip lots of checks if pv is
enabled.
IO instructions are already using logical addressing

> 
> void s390_cpu_physical_memory_write(...)
> {
>     if (!env->pv) {
>         cpu_physical_memory_write(...);
>     } else {
>         s390_cpu_virt_mem_write(...);
>     }
> }
> 
> ?
> 
>  Thomas
>