Re: [Xen-devel] [PATCH 2/2] x86emul: consolidate string insn address increments

["Jan Beulich" <JBeulich@xxxxxxxx>]

>>> On 09.12.16 at 18:11, <andrew.cooper3@xxxxxxxxxx> wrote:
> On 09/12/16 15:22, Jan Beulich wrote:
>> Move the looking at EFLAGS.DF into the macro, rendering all call sites
>> more readable.
>>
>> Signed-off-by: Jan Beulich <jbeulich@xxxxxxxx>
> 
> The net change is ok; it is certainly cleaner to read in the body of
> x86_emulate().
> 
> However, the naming of register_address_increment() was previously ok,
> as it was obvious at the calling point that a negative increment was
> possible.  This subtly is now hidden.
> 
> How about reg_addr_adjust() or reg_addr_adjust_dir() as an alternative
> name?  This retains the property that it is obvious that the direction
> flag is followed in the calculation.

That's fine with me; I admit that I've overlooked that the use of
"increment" in the name has now become further disconnected
from the actual operation. In line with
_register_address_increment() I'd prefer to call it
_register_address_adjust() though; in fact it might be reasonable
to also replace "increment" with "adjust" in the other one (as it's
similarly used for adjustments in both directions). What do you
think?

Jan


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