Hi,
El Thu, Oct 27, 2016 at 07:45:32PM +0200 H. Nikolaus Schaller ha dit:
> > Am 27.10.2016 um 19:41 schrieb Matthias Kaehlcke <m...@chromium.org>:
> >
> > El Thu, Oct 27, 2016 at 02:31:39PM +0200 H. Nikolaus Schaller ha dit:
> >
> >> commit 73e705bf81ce ("regulator: core: Add set_voltage_time op")
> >>
> >> introduced a new rdev_warn() if the ramp_delay is 0.
> >>
> >> Apparently, on omap3/twl4030 platforms with dynamic voltage
> >> management this results in non-ending spurious messages like
> >>
> >> [ 511.143066] VDD1: ramp_delay not set
> >> [ 511.662322] VDD1: ramp_delay not set
> >> [ 513.903625] VDD1: ramp_delay not set
> >> [ 514.222198] VDD1: ramp_delay not set
> >> [ 517.062835] VDD1: ramp_delay not set
> >> [ 517.382568] VDD1: ramp_delay not set
> >> [ 520.142791] VDD1: ramp_delay not set
> >> [ 520.502593] VDD1: ramp_delay not set
> >> [ 523.062896] VDD1: ramp_delay not set
> >> [ 523.362701] VDD1: ramp_delay not set
> >> [ 526.143035] VDD1: ramp_delay not set
> >>
> >> I have observed this on GTA04 while it is reported to occur on
> >> N900 as well: https://bugzilla.kernel.org/show_bug.cgi?id=178371
> >>
> >> This patch makes the warning appear only in debugging mode.
> >>
> >> Signed-off-by: H. Nikolaus Schaller <h...@goldelico.com>
> >> ---
> >> drivers/regulator/core.c | 2 +-
> >> 1 file changed, 1 insertion(+), 1 deletion(-)
> >>
> >> diff --git a/drivers/regulator/core.c b/drivers/regulator/core.c
> >> index 67426c0..5c1519b 100644
> >> --- a/drivers/regulator/core.c
> >> +++ b/drivers/regulator/core.c
> >> @@ -2754,7 +2754,7 @@ static int _regulator_set_voltage_time(struct
> >> regulator_dev *rdev,
> >> ramp_delay = rdev->desc->ramp_delay;
> >>
> >> if (ramp_delay == 0) {
> >> - rdev_warn(rdev, "ramp_delay not set\n");
> >> + rdev_dbg(rdev, "ramp_delay not set\n");
> >> return 0;
> >> }
> >
> > Sorry about that, I didn't run into this since in my tests ramp_delay
> > was set :(
>
> It may be that we have a rare exception. I didn't see it for example
> on OMAP5 based boards.
>
> >
> > Calling _regulator_set_voltage_time() with ramp_delay being zero is
> > a common case, I think the message should be removed completely.
>
> I'd have no problem if it is a debug feature, that is why I proposed
> to not completely remove it unconditionally.
It shouldn't have been added in the first place, it stems from code
that incorrectly assumes that the code path is only executed when
ramp_delay is set. I removed the message from that code
(d89564efe79419a093e966a959bf5ba2c94e693f "regulator: core: Don't skip
set_voltage_time when ramp delay disabled"), but missed to remove/not
add it in this function.
