Re: [PATCH v4 2/3] driver core: add deferring probe reason to devices_deferred property

[Andrzej Hajda]

On 20.12.2018 12:12, Greg Kroah-Hartman wrote:
> On Thu, Dec 20, 2018 at 11:22:46AM +0100, Andrzej Hajda wrote:
>> /sys/kernel/debug/devices_deferred property contains list of deferred devices.
>> This list does not contain reason why the driver deferred probe, the patch
>> improves it.
>> The natural place to set the reason is probe_err function introduced recently,
>> ie. if probe_err will be called with -EPROBE_DEFER instead of printk the message
>> will be attached to deferred device and printed when user read devices_deferred
>> property.
>>
>> Signed-off-by: Andrzej Hajda <a.hajda@xxxxxxxxxxx>
>> Reviewed-by: Mark Brown <broonie@xxxxxxxxxx>
>> Reviewed-by: Javier Martinez Canillas <javierm@xxxxxxxxxx>
>> Reviewed-by: Andy Shevchenko <andy.shevchenko@xxxxxxxxx>
>> ---
>> v4:
>>   - removed NULL check before kfree,
>>   - coding style tweaking.
>> v3:
>>   - adjusted deferred_devs_show, to accept newline ended messages,
>>   - changed conditonal check to positive,
>>   - added R-b by Andy.
>> v2:
>>   - changed __deferred_probe_set_msg args - like in __dev_printk, fits better,
>>   - use kasprintf instead of bunch of code,
>>   - keep consistent format of devices_deferred lines,
>>   - added R-Bs (again I hope changes above are not against it).
>> ---
>> ---
>>  drivers/base/base.h |  3 +++
>>  drivers/base/core.c |  9 +++++----
>>  drivers/base/dd.c   | 21 ++++++++++++++++++++-
>>  3 files changed, 28 insertions(+), 5 deletions(-)
>>
>> diff --git a/drivers/base/base.h b/drivers/base/base.h
>> index 7a419a7a6235..effbd5e7f9f1 100644
>> --- a/drivers/base/base.h
>> +++ b/drivers/base/base.h
>> @@ -75,6 +75,7 @@ struct device_private {
>>  	struct klist_node knode_driver;
>>  	struct klist_node knode_bus;
>>  	struct list_head deferred_probe;
>> +	char *deferred_probe_msg;
>>  	struct device *device;
>>  };
>>  #define to_device_private_parent(obj)	\
>> @@ -113,6 +114,8 @@ extern void device_release_driver_internal(struct device *dev,
>>  extern void driver_detach(struct device_driver *drv);
>>  extern int driver_probe_device(struct device_driver *drv, struct device *dev);
>>  extern void driver_deferred_probe_del(struct device *dev);
>> +extern void __deferred_probe_set_msg(const struct device *dev,
>> +				     struct va_format *vaf);
>>  static inline int driver_match_device(struct device_driver *drv,
>>  				      struct device *dev)
>>  {
>> diff --git a/drivers/base/core.c b/drivers/base/core.c
>> index 7f644f3c41d3..d3eb5aeeaa28 100644
>> --- a/drivers/base/core.c
>> +++ b/drivers/base/core.c
>> @@ -3108,6 +3108,7 @@ define_dev_printk_level(_dev_info, KERN_INFO);
>>   *
>>   * This helper implements common pattern present in probe functions for error
>>   * checking: print message if the error is not -EPROBE_DEFER and propagate it.
>> + * In case of -EPROBE_DEFER it sets defer probe reason.
>>   * It replaces code sequence:
>>   * 	if (err != -EPROBE_DEFER)
>>   * 		dev_err(dev, ...);
>> @@ -3123,14 +3124,14 @@ int probe_err(const struct device *dev, int err, const char *fmt, ...)
>>  	struct va_format vaf;
>>  	va_list args;
>>  
>> -	if (err == -EPROBE_DEFER)
>> -		return err;
>> -
>>  	va_start(args, fmt);
>>  	vaf.fmt = fmt;
>>  	vaf.va = &args;
>>  
>> -	dev_err(dev, "error %d: %pV", err, &vaf);
>> +	if (err == -EPROBE_DEFER)
>> +		__deferred_probe_set_msg(dev, &vaf);
>> +	else
>> +		dev_err(dev, "error %d: %pV", err, &vaf);
>>  
>>  	va_end(args);
>>  
>> diff --git a/drivers/base/dd.c b/drivers/base/dd.c
>> index 88713f182086..857aa4d1d45d 100644
>> --- a/drivers/base/dd.c
>> +++ b/drivers/base/dd.c
>> @@ -27,6 +27,7 @@
>>  #include <linux/async.h>
>>  #include <linux/pm_runtime.h>
>>  #include <linux/pinctrl/devinfo.h>
>> +#include <linux/slab.h>
>>  
>>  #include "base.h"
>>  #include "power/power.h"
>> @@ -132,6 +133,8 @@ void driver_deferred_probe_del(struct device *dev)
>>  	if (!list_empty(&dev->p->deferred_probe)) {
>>  		dev_dbg(dev, "Removed from deferred list\n");
>>  		list_del_init(&dev->p->deferred_probe);
>> +		kfree(dev->p->deferred_probe_msg);
>> +		dev->p->deferred_probe_msg = NULL;
>>  	}
>>  	mutex_unlock(&deferred_probe_mutex);
>>  }
>> @@ -202,6 +205,21 @@ void device_unblock_probing(void)
>>  	driver_deferred_probe_trigger();
>>  }
>>  
>> +/*
>> + * __deferred_probe_set_msg() - Set defer probe reason message for device
>> + */
>> +void __deferred_probe_set_msg(const struct device *dev, struct va_format *vaf)
>> +{
>> +	const char *drv = dev_driver_string(dev);
>> +
>> +	mutex_lock(&deferred_probe_mutex);
>> +
>> +	kfree(dev->p->deferred_probe_msg);
>> +	dev->p->deferred_probe_msg = kasprintf(GFP_KERNEL, "%s: %pV", drv, vaf);
> Why do you also need the dev driver string here?  Don't you get it
> automatically when you print out the list in deferred_devs_show()?


Deferred device is not bound to driver.


>
> How about we just wait on this patch, it seems very unneeded.  Or at
> least I can't see who would need this, what can a user do with this
> information?


It should be quite helpful in case of deferred probing, solving
mysteries why given device is not bound to the driver is very common in
mobile world.

And since deferal should not be treated as an error grepping dmesg often
does not help.


Regards

Andrzej



>
> thanks,
>
> greg k-h
>
>