Re: [PATCH v2 3/4] migration/calc-dirty-rate: added n-zero-pages metric

[Juan Quintela <quintela@redhat.com>]

Andrei Gudkov <gudkov.andrei@huawei.com> wrote:
> In sampling mode, a new metric is collected and reported:
> number of pages entirely filled with zeroes.

Good idea.

> @@ -331,11 +336,20 @@ static uint32_t compute_page_hash(void *ptr)
>      v2 = QEMU_XXHASH_SEED + XXH_PRIME64_2;
>      v3 = QEMU_XXHASH_SEED + 0;
>      v4 = QEMU_XXHASH_SEED - XXH_PRIME64_1;
> -    for (i = 0; i < TARGET_PAGE_SIZE / 8; i += 4) {
> -        v1 = XXH64_round(v1, p[i + 0]);
> -        v2 = XXH64_round(v2, p[i + 1]);
> -        v3 = XXH64_round(v3, p[i + 2]);
> -        v4 = XXH64_round(v4, p[i + 3]);
> +    if (ptr) {

It smells like a hack, that is only going to be used once in the
program.
But the only other option that I can think is repeating the function for
the zero case.

No way to win here.


> +        for (i = 0; i < TARGET_PAGE_SIZE / 8; i += 4) {
> +            v1 = XXH64_round(v1, p[i + 0]);
> +            v2 = XXH64_round(v2, p[i + 1]);
> +            v3 = XXH64_round(v3, p[i + 2]);
> +            v4 = XXH64_round(v4, p[i + 3]);
> +        }
> +    } else {
> +        for (i = 0; i < TARGET_PAGE_SIZE / 8; i += 4) {
> +            v1 = XXH64_round(v1, 0);
> +            v2 = XXH64_round(v2, 0);
> +            v3 = XXH64_round(v3, 0);
> +            v4 = XXH64_round(v4, 0);
> +        }

>      }
>      res = XXH64_mergerounds(v1, v2, v3, v4);
>      res += TARGET_PAGE_SIZE;
> @@ -343,6 +357,17 @@ static uint32_t compute_page_hash(void *ptr)
>      return (uint32_t)(res & UINT32_MAX);
>  }
>  
> +static uint32_t get_zero_page_hash(void)
> +{
> +    static uint32_t hash;
> +    static int is_computed;

bool?


Later, Juan.