Re: [PATCH] Input: silead: use msleep() for long delays

[Nicholas Mc Guire]

On Thu, Jan 12, 2017 at 10:10:44AM -0800, Dmitry Torokhov wrote:
> On Thu, Jan 12, 2017 at 9:46 AM, Hans de Goede <hdegoede@xxxxxxxxxx> wrote:
> > Hi,
> >
> > On 01/12/2017 05:21 PM, Nicholas Mc Guire wrote:
> >>
> >> the delays here are in the 10 to 20ms range so msleep() will do - no
> >> need to burden the highres timer subsystem.
> >>
> >> Signed-off-by: Nicholas Mc Guire <hofrat@xxxxxxxxx>
> >> ---
> >>
> >> Problem found by coccinelle script
> >>
> >> While msleep(10) has a worst case uncertainty of 10ms (on HZ=100 systems)
> >> this seems ok here as the delays are not called frequently (init and
> >> reset functions)
> >
> >
> > By the same logic, this is not much of a burden on the high-res timer
> > subsys though.
> >
> >> and the uncertainty of 10ms fits the permitted range of
> >> the original usleep_ranges().
> >
> >
> > Either way this patch is fine with me.
> 
> I'd rather not because next will come a checkpatch warrior and I will
> have to convince them why msleep is OK here. And another one, and
> another one... :(
>
there is no checkpatch warning here - checkpatch only throws warnings
of range < 20ms if hardcoded but this is in a #define so its fine with
respect to checkpatch.

But if there are concerns with this - thats fine - its most likely not
critical - the goal is to have a consistent usage of highres timers - 
including limiting there use to the cases where its really needed.

thx!
hofrat