Drop the redundant probe debug message which is already provided by
driver core. Whether probe succeeded can also be determined through
sysfs.
Signed-off-by: Johan Hovold <[email protected]>
---
drivers/soundwire/bus_type.c | 2 --
1 file changed, 2 deletions(-)
diff --git a/drivers/soundwire/bus_type.c b/drivers/soundwire/bus_type.c
index 43e732ca73c0..5b9b129badff 100644
--- a/drivers/soundwire/bus_type.c
+++ b/drivers/soundwire/bus_type.c
@@ -152,8 +152,6 @@ static int sdw_drv_probe(struct device *dev)
mutex_unlock(&slave->sdw_dev_lock);
- dev_dbg(dev, "probe complete\n");
-
return 0;
}
--
2.44.1
On 6/4/24 02:52, Johan Hovold wrote:
> Drop the redundant probe debug message which is already provided by
> driver core. Whether probe succeeded can also be determined through
> sysfs.
>
> Signed-off-by: Johan Hovold <[email protected]>
> ---
> drivers/soundwire/bus_type.c | 2 --
> 1 file changed, 2 deletions(-)
>
> diff --git a/drivers/soundwire/bus_type.c b/drivers/soundwire/bus_type.c
> index 43e732ca73c0..5b9b129badff 100644
> --- a/drivers/soundwire/bus_type.c
> +++ b/drivers/soundwire/bus_type.c
> @@ -152,8 +152,6 @@ static int sdw_drv_probe(struct device *dev)
>
> mutex_unlock(&slave->sdw_dev_lock);
>
> - dev_dbg(dev, "probe complete\n");
> -
> return 0;
> }
I don't see the point of removing this, we've used it for the last 5
years to figure out when the probe complete vs. when the device becomes
attached. It's a simple log that helped us immensely with race
conditions, etc.
And no, sysfs is not a viable alternative to look at CI logs, and
enabling dynamic debug on the driver core would be way to verbose.
On Tue, Jun 04, 2024 at 10:37:17AM +0200, Pierre-Louis Bossart wrote:
> On 6/4/24 02:52, Johan Hovold wrote:
> > Drop the redundant probe debug message which is already provided by
> > driver core. Whether probe succeeded can also be determined through
> > sysfs.
> > @@ -152,8 +152,6 @@ static int sdw_drv_probe(struct device *dev)
> >
> > mutex_unlock(&slave->sdw_dev_lock);
> >
> > - dev_dbg(dev, "probe complete\n");
> > -
> > return 0;
> > }
>
> I don't see the point of removing this, we've used it for the last 5
> years to figure out when the probe complete vs. when the device becomes
> attached. It's a simple log that helped us immensely with race
> conditions, etc.
Fair enough. Soundwire probing is indeed a bit of a mess.
Johan