Re: [x86 setup 17/33] A20 handling code

["H. Peter Anvin" <hpa@zytor.com>]

Jan Engelhardt wrote:
>> +
>> +#define A20_TEST_ADDR	(4*0x80)
>> +#define A20_TEST_SHORT  32
>> +#define A20_TEST_LONG	2097152	/* 2^21 */
> 
> Maybe...?
> #define A20_TEST_LONG (1 << 21)

That makes it look like it's a magic value or bitmask, it's not.  The
value 2^21 is largely arbitrary, it's just what happened to be in the
previous code.

>> +/* Quick test to see if A20 is already enabled */
>> +static int a20_test_short(void)
>> +{
>> +	return a20_test(A20_TEST_SHORT);
>> +}
>> +
>> +/* Longer test that actually waits for A20 to come on line; this
>> +   is useful when dealing with the KBC or other slow external circuitry. */
>> +static int a20_test_long(void)
>> +{
>> +	return a20_test(A20_TEST_LONG);
>> +}
> 
> To me looks like some of these (or other functions) could return bool.

Does it matter?  It will generate worse code.

	-hpa