On 31/03/17 03:47, Arushi Singhal wrote: > The driver needs to insure atomicity during frequency > changes of bus and device. The iiodev->mlock as used > was not doing that. Replace it with the drivers existing > buffer lock and execute spi_write directly. > > Signed-off-by: Arushi Singhal <arushisinghal19971997@xxxxxxxxx> Firstly please put the driver name in the title when a patch only effects one driver. Secondly, please see final version of Gargi Sharma's patch for the ade7754. As with that one, the basic principle is fine, but I didn't like the V2 version which did the local copy of spi_write as you have here. Introduced an unlocked version of the ade7759_spi_write_reg_16 function and call it both from ade7759_spi_write_reg_16 and directly here. Puts the _WRITE_REG magic all in one place. Otherwise, looks good! Jonathan > --- > drivers/staging/iio/meter/ade7759.c | 9 ++++++--- > 1 file changed, 6 insertions(+), 3 deletions(-) > > diff --git a/drivers/staging/iio/meter/ade7759.c b/drivers/staging/iio/meter/ade7759.c > index 0b65f1847510..a5ee4501d30a 100644 > --- a/drivers/staging/iio/meter/ade7759.c > +++ b/drivers/staging/iio/meter/ade7759.c > @@ -429,7 +429,7 @@ static ssize_t ade7759_write_frequency(struct device *dev, > if (!val) > return -EINVAL; > > - mutex_lock(&indio_dev->mlock); > + mutex_lock(&st->buf_lock); > > t = 27900 / val; > if (t > 0) > @@ -447,10 +447,13 @@ static ssize_t ade7759_write_frequency(struct device *dev, > reg &= ~(3 << 13); > reg |= t << 13; > > - ret = ade7759_spi_write_reg_16(dev, ADE7759_MODE, reg); > + st->tx[0] = ADE7759_WRITE_REG(ADE7759_MODE); > + st->tx[1] = reg; > + > + ret = spi_write(st->us, st->tx, 2); > > out: > - mutex_unlock(&indio_dev->mlock); > + mutex_unlock(&st->buf_lock); > > return ret ? ret : len; > } > _______________________________________________ devel mailing list devel@xxxxxxxxxxxxxxxxxxxxxx http://driverdev.linuxdriverproject.org/mailman/listinfo/driverdev-devel
