2014-01-02 02:02:10

by Zhang, Rui

Subject: Re: [PATCH V2] thermal: imx: correct suspend/resume flow

On Tue, 2013-12-24 at 09:43 -0500, Anson Huang wrote:
> Fixes regression introduced by:
>
> commit 37713a1e8e4c1a1067ad4c99296f78d3c82ed9c4
> Author: Philipp Zabel <[email protected]>
> Date: Thu Aug 1 18:33:12 2013 +0200
>
> thermal: imx: implement thermal alarm interrupt handling
>
> The commit 37713a1e8e4 makes imx thermal sensor always powered up as alarm
> function is enabled, but the suspend callback of imx thermal returns
> success only if thermal sensor is powered down, so it will always returns
> fail hence break system's suspend, this patch disables imx thermal sensor
> before suspend and re-enable it after resume.
>
> Signed-off-by: Anson Huang <[email protected]>

applied. BTW, this should also be a stable material for 3.12, right?

thanks,
rui
> ---
> drivers/thermal/imx_thermal.c | 31 +++++++++++++++++--------------
> 1 file changed, 17 insertions(+), 14 deletions(-)
>
> diff --git a/drivers/thermal/imx_thermal.c b/drivers/thermal/imx_thermal.c
> index cbb16f3..9a9a6c2 100644
> --- a/drivers/thermal/imx_thermal.c
> +++ b/drivers/thermal/imx_thermal.c
> @@ -510,27 +510,30 @@ static int imx_thermal_suspend(struct device *dev)
> {
> struct imx_thermal_data *data = dev_get_drvdata(dev);
> struct regmap *map = data->tempmon;
> - u32 val;
>
> - regmap_read(map, TEMPSENSE0, &val);
> - if ((val & TEMPSENSE0_POWER_DOWN) == 0) {
> - /*
> - * If a measurement is taking place, wait for a long enough
> - * time for it to finish, and then check again. If it still
> - * does not finish, something must go wrong.
> - */
> - udelay(50);
> - regmap_read(map, TEMPSENSE0, &val);
> - if ((val & TEMPSENSE0_POWER_DOWN) == 0)
> - return -ETIMEDOUT;
> - }
> + /*
> + * Need to disable thermal sensor, otherwise, when thermal core
> + * try to get temperature before thermal sensor resume, a wrong
> + * temperature will be read as the thermal sensor is powered
> + * down.
> + */
> + regmap_write(map, TEMPSENSE0 + REG_CLR, TEMPSENSE0_MEASURE_TEMP);
> + regmap_write(map, TEMPSENSE0 + REG_SET, TEMPSENSE0_POWER_DOWN);
> + data->mode = THERMAL_DEVICE_DISABLED;
>
> return 0;
> }
>
> static int imx_thermal_resume(struct device *dev)
> {
> - /* Nothing to do for now */
> + struct imx_thermal_data *data = dev_get_drvdata(dev);
> + struct regmap *map = data->tempmon;
> +
> + /* Enabled thermal sensor after resume */
> + regmap_write(map, TEMPSENSE0 + REG_CLR, TEMPSENSE0_POWER_DOWN);
> + regmap_write(map, TEMPSENSE0 + REG_SET, TEMPSENSE0_MEASURE_TEMP);
> + data->mode = THERMAL_DEVICE_ENABLED;
> +
> return 0;
> }
> #endif

2014-01-02 02:16:56

by Shawn Guo

Subject: Re: [PATCH V2] thermal: imx: correct suspend/resume flow

On Thu, Jan 02, 2014 at 10:01:51AM +0800, Zhang Rui wrote:
> On Tue, 2013-12-24 at 09:43 -0500, Anson Huang wrote:
> > Fixes regression introduced by:
> >
> > commit 37713a1e8e4c1a1067ad4c99296f78d3c82ed9c4
> > Author: Philipp Zabel <[email protected]>
> > Date: Thu Aug 1 18:33:12 2013 +0200
> >
> > thermal: imx: implement thermal alarm interrupt handling
> >
> > The commit 37713a1e8e4 makes imx thermal sensor always powered up as alarm
> > function is enabled, but the suspend callback of imx thermal returns
> > success only if thermal sensor is powered down, so it will always returns
> > fail hence break system's suspend, this patch disables imx thermal sensor
> > before suspend and re-enable it after resume.
> >
> > Signed-off-by: Anson Huang <[email protected]>
>
> applied. BTW, this should also be a stable material for 3.12, right?

Right. That would be great. Thanks.

Shawn