Re: [PATCH 4/5] iio: imu: st_lsm6dsx: always enter interrupt thread

[Jonathan Cameron <jic23@kernel.org>]

On Tue, 18 Jun 2019 22:24:14 +0200
Lorenzo Bianconi <lorenzo@kernel.org> wrote:

> On Jun 18, Sean Nyekjaer wrote:
> > 
> > 
> > On 18/06/2019 17.55, Lorenzo Bianconi wrote:  
> > > > The interrupt source can come from multiple sources, fifo and wake interrupts.
> > > > Enter interrupt thread to check which interrupt that has fired.
> > > > 
> > > > Signed-off-by: Sean Nyekjaer <sean@geanix.com>
> > > > ---
> > > >   drivers/iio/imu/st_lsm6dsx/st_lsm6dsx_core.c | 30 +++++++++++++++-----
> > > >   1 file changed, 23 insertions(+), 7 deletions(-)
> > > > 
> > > > diff --git a/drivers/iio/imu/st_lsm6dsx/st_lsm6dsx_core.c b/drivers/iio/imu/st_lsm6dsx/st_lsm6dsx_core.c
> > > > index 59a34894e495..76aec5024d83 100644
> > > > --- a/drivers/iio/imu/st_lsm6dsx/st_lsm6dsx_core.c
> > > > +++ b/drivers/iio/imu/st_lsm6dsx/st_lsm6dsx_core.c
> > > > @@ -78,6 +78,12 @@
> > > >   #define ST_LSM6DSX_REG_GYRO_OUT_Y_L_ADDR	0x24
> > > >   #define ST_LSM6DSX_REG_GYRO_OUT_Z_L_ADDR	0x26
> > > > +#define ST_LSM6DSX_REG_WAKE_UP_SRC_ADDR		0x1B
> > > > +#define ST_LSM6DSX_REG_WAKE_UP_SRC_Z_WU_MASK	BIT(0)
> > > > +#define ST_LSM6DSX_REG_WAKE_UP_SRC_Y_WU_MASK	BIT(1)
> > > > +#define ST_LSM6DSX_REG_WAKE_UP_SRC_X_WU_MASK	BIT(2)
> > > > +#define ST_LSM6DSX_REG_WAKE_UP_SRC_WU_MASK	BIT(4)
> > > > +
> > > >   #define ST_LSM6DSX_REG_TAP_CFG_ADDR		0x58
> > > >   #define ST_LSM6DSX_REG_TAP_CFG_INT_EN_MASK	BIT(7)
> > > >   #define ST_LSM6DSX_REG_TAP_CFG_INACT_EN_MASK	GENMASK(6, 5)
> > > > @@ -946,19 +952,29 @@ int st_lsm6dsx_event_setup(struct st_lsm6dsx_hw *hw)
> > > >   static irqreturn_t st_lsm6dsx_handler_irq(int irq, void *private)
> > > >   {
> > > > -	struct st_lsm6dsx_hw *hw = private;
> > > > -
> > > > -	return hw->sip > 0 ? IRQ_WAKE_THREAD : IRQ_NONE;
> > > > +	return IRQ_WAKE_THREAD;  
> > > 
> > > I guess this will break shared interrupt, isn't it?  
> > 
> > When you write shared interrupt you mean: the drdy-int-pin stuff?
> > I need to add so we can use all this functionality with the INT2 as well...  
> 
> nope, shared IRQ line with other devices (when you set drive-open-drain dts
> property)

It's been a while since I looked at this, but...

It shouldn't be broken.  When using shared interrupts, all interrupt handlers
tend to get run, whether or not a given one return IRQ_NONE.

Nothing stops multiple devices setting their interrupt lines at the same
time.

See __handle_irq_event_percpu in kernel/irq/handle.c which is probably
the right code. No return values are taken notice of until all registered
handlers have run.

Jonathan

> 
> >   
> > >   
> > > >   }
> > > >   static irqreturn_t st_lsm6dsx_handler_thread(int irq, void *private)
> > > >   {
> > > >   	struct st_lsm6dsx_hw *hw = private;
> > > > -	int count;
> > > > +	int count = 0;
> > > > +	int data, err;
> > > > +
> > > > +	if (hw->enable_event) {
> > > > +		err = regmap_read(hw->regmap,
> > > > +				  ST_LSM6DSX_REG_WAKE_UP_SRC_ADDR, &data);
> > > > +		if (err < 0)
> > > > +			goto try_fifo;  
> > > 
> > > You can simplify this just doing something like:
> > > 
> > > 		if (err < 0 && !hw->sip)
> > > 			return IRQ_NONE;
> > >   
> > Will apply :-)
> > 
> > Thanks for the review Lorenzo...  
> 
> Thanks for working on this :)
> 
> Regards,
> Lorenzo
> 
> > 
> > -- 
> > Best regards,
> > Sean Nyekjær
> > Embedded Linux Consultant
> > 
> > +45 42427326
> > sean@geanix.com
> > 
> > Geanix ApS
> > https://geanix.com
> > DK39600706