Re: [Xen-devel] [PATCH] x86/HVM: handle_{mmio*, pio}() return value adjustments

["Jan Beulich" <JBeulich@xxxxxxxx>]

>>> On 16.12.16 at 10:52, <Paul.Durrant@xxxxxxxxxx> wrote:
>> From: Jan Beulich [mailto:JBeulich@xxxxxxxx]
>> Sent: 16 December 2016 09:31
>> @@ -100,22 +100,21 @@ int handle_mmio(void)
>>      {
>>      case X86EMUL_UNHANDLEABLE:
>>          hvm_dump_emulation_state(XENLOG_G_WARNING "MMIO", &ctxt);
>> -        return 0;
>> +        return false;
>> +
>>      case X86EMUL_EXCEPTION:
>>          if ( ctxt.ctxt.event_pending )
>>              hvm_inject_event(&ctxt.ctxt.event);
>>          break;
>> -    default:
>> -        break;
> 
> Should there not be some sort of default case, even if it's simply to assert 
> that it's not reachable?

A default case doing nothing when the switch expression is not of
an enum type is pointless. And it _is_ reachable (namely for
X86EMUL_OKAY).

>> --- a/xen/arch/x86/hvm/ioreq.c
>> +++ b/xen/arch/x86/hvm/ioreq.c
>> @@ -156,13 +156,14 @@ bool_t handle_hvm_io_completion(struct v
> 
> Do you not want to change this to from bool_t to bool while you're at it?

Well, while I first was inclined to do so, I then didn't want to do
too many things at once (iirc there would be a few more which then
would want updating at the same time).

Jan


_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxx
https://lists.xen.org/xen-devel