Re: [PATCH v3 08/18] target/s390: Return exception from mmu_translate_real

[David Hildenbrand]

On 27.09.19 12:44, David Hildenbrand wrote:
> On 26.09.19 18:26, Richard Henderson wrote:
>> Do not raise the exception directly within mmu_translate_real,
>> but pass it back so that caller may do so.
>>
>> Signed-off-by: Richard Henderson <address@hidden>
>> ---
>>  target/s390x/internal.h    | 2 +-
>>  target/s390x/excp_helper.c | 4 ++--
>>  target/s390x/mmu_helper.c  | 8 ++++----
>>  3 files changed, 7 insertions(+), 7 deletions(-)
>>
>> diff --git a/target/s390x/internal.h b/target/s390x/internal.h
>> index c243fa725b..c4388aaf23 100644
>> --- a/target/s390x/internal.h
>> +++ b/target/s390x/internal.h
>> @@ -362,7 +362,7 @@ void probe_write_access(CPUS390XState *env, uint64_t 
>> addr, uint64_t len,
>>  int mmu_translate(CPUS390XState *env, target_ulong vaddr, int rw, uint64_t 
>> asc,
>>                    target_ulong *raddr, int *flags, bool exc);
>>  int mmu_translate_real(CPUS390XState *env, target_ulong raddr, int rw,
>> -                       target_ulong *addr, int *flags);
>> +                       target_ulong *addr, int *flags, uint64_t *tec);
>>  
>>  
>>  /* misc_helper.c */
>> diff --git a/target/s390x/excp_helper.c b/target/s390x/excp_helper.c
>> index ab2ed47fef..906b87c071 100644
>> --- a/target/s390x/excp_helper.c
>> +++ b/target/s390x/excp_helper.c
>> @@ -147,8 +147,8 @@ bool s390_cpu_tlb_fill(CPUState *cs, vaddr address, int 
>> size,
>>          if (!(env->psw.mask & PSW_MASK_64)) {
>>              vaddr &= 0x7fffffff;
>>          }
>> -        fail = mmu_translate_real(env, vaddr, access_type, &raddr, &prot);
>> -        excp = 0; /* exception already raised */
>> +        excp = mmu_translate_real(env, vaddr, access_type, &raddr, &prot, 
>> &tec);
>> +        fail = excp;
>>      } else {
>>          g_assert_not_reached();
>>      }
>> diff --git a/target/s390x/mmu_helper.c b/target/s390x/mmu_helper.c
>> index 3ef40a37a7..b783c62bd7 100644
>> --- a/target/s390x/mmu_helper.c
>> +++ b/target/s390x/mmu_helper.c
>> @@ -523,10 +523,10 @@ void s390_cpu_virt_mem_handle_exc(S390CPU *cpu, 
>> uintptr_t ra)
>>   * @param rw     0 = read, 1 = write, 2 = code fetch
>>   * @param addr   the translated address is stored to this pointer
>>   * @param flags  the PAGE_READ/WRITE/EXEC flags are stored to this pointer
>> - * @return       0 if the translation was successful, < 0 if a fault 
>> occurred
>> + * @return       0 = success, != 0, the exception to raise
>>   */
>>  int mmu_translate_real(CPUS390XState *env, target_ulong raddr, int rw,
>> -                       target_ulong *addr, int *flags)
>> +                       target_ulong *addr, int *flags, uint64_t *tec)
>>  {
>>      const bool lowprot_enabled = env->cregs[0] & CR0_LOWPROT;
>>  
>> @@ -535,8 +535,8 @@ int mmu_translate_real(CPUS390XState *env, target_ulong 
>> raddr, int rw,
>>          /* see comment in mmu_translate() how this works */
>>          *flags |= PAGE_WRITE_INV;
>>          if (is_low_address(raddr) && rw == MMU_DATA_STORE) {
>> -            trigger_access_exception(env, PGM_PROTECTION, ILEN_AUTO, 0);
>> -            return -EACCES;
>> +            *tec = 0;
>> +            return PGM_PROTECTION;
>>          }
>>      }
>>  
>>
> 
> Note that
> 
> [PATCH v1 2/5] s390x/mmu: Implement ESOP-2 and
> access-exception-fetch/store-indication facility
> 
> also messes with the tec (which is okay), but also with the ILEN on
> instruction fetches.

I'll drop the ilen change from that patch, now that I figured out how it
works :)

-- 

Thanks,

David / dhildenb