On 5/20/19 6:17 PM, Darrick J. Wong wrote:
> From: Darrick J. Wong <darrick.wong@xxxxxxxxxx>
>
> The comment preceding background_sleep() is wrong -- the function sleeps
> 100us, not 100ms, for every '-b' passed in after the first one. This is
> really not obvious from the magic numbers, so fix the comment and use
> symbolic constants for easier reading.
>
> Signed-off-by: Darrick J. Wong <darrick.wong@xxxxxxxxxx>
Ok, looks like man page was already correct.
Reviewed-by: Eric Sandeen <sandeen@xxxxxxxxxx>
> ---
> scrub/common.c | 12 +++++++-----
> 1 file changed, 7 insertions(+), 5 deletions(-)
>
>
> diff --git a/scrub/common.c b/scrub/common.c
> index c877c7c8..1cd2b7ba 100644
> --- a/scrub/common.c
> +++ b/scrub/common.c
> @@ -253,21 +253,23 @@ scrub_nproc_workqueue(
> }
>
> /*
> - * Sleep for 100ms * however many -b we got past the initial one.
> + * Sleep for 100us * however many -b we got past the initial one.
> * This is an (albeit clumsy) way to throttle scrub activity.
> */
> +#define NSEC_PER_SEC 1000000000ULL
> +#define NSEC_PER_USEC 1000ULL
> void
> background_sleep(void)
> {
> - unsigned long long time;
> + unsigned long long time_ns;
> struct timespec tv;
>
> if (bg_mode < 2)
> return;
>
> - time = 100000ULL * (bg_mode - 1);
> - tv.tv_sec = time / 1000000;
> - tv.tv_nsec = time % 1000000;
> + time_ns = 100 * NSEC_PER_USEC * (bg_mode - 1);
> + tv.tv_sec = time_ns / NSEC_PER_SEC;
> + tv.tv_nsec = time_ns % NSEC_PER_SEC;
> nanosleep(&tv, NULL);
> }
>
>
