Hi
On Sun, Apr 20, 2014 at 12:05 AM, Pavel Machek <pavel@xxxxxxx> wrote:
>> +static void pgd_init(pgd_t *pgd)
>> +{
>> + unsigned long *p = (unsigned long *) pgd;
>> + int i;
>> +
>> + for (i = 0; i < USER_PTRS_PER_PGD; i += 8) {
>> + p[i + 0] = (unsigned long) invalid_pte_table;
>> + p[i + 1] = (unsigned long) invalid_pte_table;
>> + p[i + 2] = (unsigned long) invalid_pte_table;
>> + p[i + 3] = (unsigned long) invalid_pte_table;
>> + p[i + 4] = (unsigned long) invalid_pte_table;
>> + p[i + 5] = (unsigned long) invalid_pte_table;
>> + p[i + 6] = (unsigned long) invalid_pte_table;
>> + p[i + 7] = (unsigned long) invalid_pte_table;
>> + }
>
> Umm. Manual loop unrolling. Does it really improve performance? Is the
> code hot enough to warrant the uglyness?
We can use the memset here because invalid_pte_table is full 32 bit value.
BTW, arch for score and mips have the similar implementation as well.
Regards
--
To unsubscribe from this list: send the line "unsubscribe linux-arch" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
