Hi, On 18-09-14 15:57, Mauro Carvalho Chehab wrote: > Em Fri, 14 Sep 2018 20:20:46 +0200 > Marco Felsch <m.felsch@xxxxxxxxxxxxxx> escreveu: > > > Hi Sakari, > > > > On 18-09-14 16:23, Sakari Ailus wrote: > > > Hi Marco, > > > > > > On Mon, Aug 13, 2018 at 11:25:08AM +0200, Marco Felsch wrote: > > > > Don't en-/disable the interrupts during s_stream because someone can > > > > disable the stream but wants to get informed if the stream is locked > > > > again. So keep the interrupts enabled the whole time the pipeline is > > > > opened. > > > > > > > > Signed-off-by: Marco Felsch <m.felsch@xxxxxxxxxxxxxx> > > > > --- > > > > drivers/media/i2c/tvp5150.c | 23 +++++++++++++++++------ > > > > 1 file changed, 17 insertions(+), 6 deletions(-) > > > > > > > > diff --git a/drivers/media/i2c/tvp5150.c b/drivers/media/i2c/tvp5150.c > > > > index e736f609fecd..e296f5bfae21 100644 > > > > --- a/drivers/media/i2c/tvp5150.c > > > > +++ b/drivers/media/i2c/tvp5150.c > > > > @@ -1389,11 +1389,26 @@ static const struct media_entity_operations tvp5150_sd_media_ops = { > > > > /**************************************************************************** > > > > I2C Command > > > > ****************************************************************************/ > > > > +static int tvp5150_s_power(struct v4l2_subdev *sd, int on) > > > > +{ > > > > + struct tvp5150 *decoder = to_tvp5150(sd); > > > > + unsigned int val = 0; > > > > + > > > > + if (on) > > > > + val = TVP5150_INT_A_LOCK; > > > > + > > > > + if (decoder->irq) > > > > + /* Enable / Disable lock interrupt */ > > > > + regmap_update_bits(decoder->regmap, TVP5150_INT_ENABLE_REG_A, > > > > + TVP5150_INT_A_LOCK, val); > > > > > > Could you use runtime PM instead? > > > > I will test it next monday. What's the different between s_power and > > runtime PM? > > > > > > > > For an example, the dw9714 driver does this: drivers/media/i2c/dw9714.c . > > > > Hopefully I got you right, should I use the > > v4l2_subdev_internal_ops.open/close and call the pm_runtime_put/get > > there or did you mean the driver.pm callbacks? I'm not that familiar > > with the pm ops at the moment, sorry. I did some tests with the PM runtime ops. I used the suspend() callback to disbale the IRQ and the resume() callback to enable it. But this won't work with my IRQ-driven setup because if the signal is not locked (maybe lost due to no input stream) the tvp went into suspend. In this mode the IRQ is disabled and I have no chance to enable the chip again. As Mauro pointed out below, there maybe some other issues I can't test at the moment. I'm with Mauro to keep the .s_power() callback for this series. There should be another patchset which covers the PM rework for all devices using the TVP5150. > > I guess the main issue here is: will this work if the bridge > driver is em28xx? > > Whatever change we do, tvp5150 should still fully work with em28xx, > as several devices use this demod there. > > Changing em28xx to cope with runtime PM would be *very* complex, > as there are lots of other drivers that can work with it, and > touching those will affect lots of other drivers. At the end, it > will very likely affect all PCI/PCIe V4L2 drivers, and several > USB ones. > > If it can be done without affecting PM with em28xx, let's do it. > Otherwise, let's stick with s_power on this series, and let > the mass PM rework on non-platform drivers to happen on some > separate patchset. > > Thanks, > Mauro > Kind regards, Marco