Messages in this thread | | | Date | Tue, 12 Jul 2016 14:24:39 -0700 | From | Guenter Roeck <> | Subject | Re: [PATCH v3 1/3] hwmon: (tmp102) Improve handling of initial read delay |
| |
On Tue, Jul 12, 2016 at 03:53:17PM -0500, Nishanth Menon wrote: > On 06/29/2016 10:41 PM, Guenter Roeck wrote: > >If the chip was in shutdown mode when the driver was loaded, the first > >conversion is ready no more than 35 milli-seconds after the chip was > >taken out of shutdown. The driver delay was so far set to 333 ms (HZ / 3), > >which is much higher than the maximum time needed by the chip. > >Reduce the time to 35 milli-seconds. > > > >Introduce a 'valid' flag to ensure that sensor data is actually read > >even if requested less than 333 ms after the driver was loaded. > > > >Signed-off-by: Guenter Roeck <linux@roeck-us.net> > >--- > >v3: Don't go to sleep if sensor readings are not available; > > instead, return -EAGAIN (matching previous behavior) > > tmp102_update_device() is now a void function. > >v2: No change > > > >Note: This series depends on the following two patches in -next. > > > a little too late perhaps.. Apologies, was on vacation.. tested and for the > series: > > Acked-by: Nishanth Menon <nm@ti.com> > Thanks!
Guenter
| |