Re: fio parsing issue when trying to use latency_target and latency_window

On Mon, Mar 16, 2026 at 11:19=E2=80=AFAM Vincent Fu <[email protected]> w=
rote:
>
> On Thu, Mar 12, 2026 at 2:46=E2=80=AFPM Mark Erdos <[email protected]>=
 wrote:
> >
> > Hello,
> >
> > With further testing I can see that the input is in microseconds both
> > for the latency target and the window, and that means the manual is
> > correct too. So, yes, if the output could match that or per Vincent,
> > your suggestion had the unit designation, it would certainly prevent
> > confusion for the next person. I am ok, since Vincent gave me the key.
> > Thanks, Mark
> >
>
> I committed a patch to the repository that adds units to the normal
> and JSON output. If you are able to test this, please report back if
> you encounter any problems.
>
> There is still a problem that --latency_target=3D10000ns specified in
> nsec units is not parsed correctly.
>
> Vincent

Upon closer investigation, fio has never supported specifying time
using nsec units, actually.

Vincent