Re: [Qemu-trivial] [Qemu-devel] [PATCH] hw/core/register: Log unimplemented access via the 'unimp' debug level

[Philippe Mathieu-Daudé]

On 06/22/2018 04:49 PM, Alistair Francis wrote:
> On Fri, Jun 22, 2018 at 5:47 AM, Philippe Mathieu-Daudé <address@hidden> 
> wrote:
>> Signed-off-by: Philippe Mathieu-Daudé <address@hidden>
>> ---
>>  hw/core/register.c | 4 ++--
>>  1 file changed, 2 insertions(+), 2 deletions(-)
>>
>> diff --git a/hw/core/register.c b/hw/core/register.c
>> index d2d1636250..8ed7c6b927 100644
>> --- a/hw/core/register.c
>> +++ b/hw/core/register.c
>> @@ -203,7 +203,7 @@ void register_write_memory(void *opaque, hwaddr addr,
>>      }
>>
>>      if (!reg) {
>> -        qemu_log_mask(LOG_GUEST_ERROR, "%s: write to unimplemented register 
>> " \
>> +        qemu_log_mask(LOG_UNIMP, "%s: write to unimplemented register " \
>>                        "at address: %#" PRIx64 "\n", reg_array->prefix, 
>> addr);
>>          return;
>>      }
>> @@ -232,7 +232,7 @@ uint64_t register_read_memory(void *opaque, hwaddr addr,
>>      }
>>
>>      if (!reg) {
>> -        qemu_log_mask(LOG_GUEST_ERROR, "%s:  read to unimplemented register 
>> " \
>> +        qemu_log_mask(LOG_UNIMP, "%s:  read to unimplemented register " \
>>                        "at address: %#" PRIx64 "\n", reg_array->prefix, 
>> addr);
> 
> I'm not sure this is correct. These can be unimplemented because there
> are gaps in the register memory map, so it is a guest error.

This is exactly what I was wondering, thanks for your comment, I'll
probably add a comment about it.

> 
> Alistair
> 
>>          return 0;
>>      }
>> --
>> 2.18.0.rc2
>>
>>