Re: [PATCH 2/2] soundwire: sysfs: add slave status and device number before probe

[Greg KH <gregkh@linuxfoundation.org>]

On Wed, Sep 16, 2020 at 03:15:04PM -0500, Pierre-Louis Bossart wrote:
> The MIPI DisCo device properties that are read by the driver from
> platform firmware, or hard-coded in the driver, should only be
> provided as sysfs entries when a driver probes successfully.
> 
> However the device status and device number is updated even when there
> is no driver present, and hence can be updated when a Slave device is
> detected on the bus without being described in platform firmware and
> without any driver registered/probed.
> 
> The attr group added for Slave status and device number is not handled
> by devm_ routines to avoid errors such as "Resources present before
> probing". Since device_del() explicitly removes attribute groups, only
> an init function is provided.
> 
> Credits to Vinod Koul for the status_show() function, shared in a
> separate patch and used as is here. The status table was modified to
> remove an unnecessary enum and status_show() is handled in a different
> group attribute than what was suggested by Vinod.
> 
> Reviewed-by: Bard Liao <yung-chuan.liao@linux.intel.com>
> Reviewed-by: Rander Wang <rander.wang@linux.intel.com>
> Reviewed-by: Guennadi Liakhovetski <guennadi.liakhovetski@linux.intel.com>
> Tested-by: Srinivas Kandgatla <srinivas.kandagatla@linaro.org>
> Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
> ---
>  .../ABI/testing/sysfs-bus-soundwire-slave     | 18 ++++++
>  drivers/soundwire/slave.c                     |  2 +
>  drivers/soundwire/sysfs_local.h               |  4 ++
>  drivers/soundwire/sysfs_slave.c               | 64 ++++++++++++++++++-
>  4 files changed, 87 insertions(+), 1 deletion(-)
> 
> diff --git a/Documentation/ABI/testing/sysfs-bus-soundwire-slave b/Documentation/ABI/testing/sysfs-bus-soundwire-slave
> index db4c9511d1aa..425adf7b8aec 100644
> --- a/Documentation/ABI/testing/sysfs-bus-soundwire-slave
> +++ b/Documentation/ABI/testing/sysfs-bus-soundwire-slave
> @@ -1,3 +1,21 @@
> +What:		/sys/bus/soundwire/devices/sdw:.../dev-status/status
> +		/sys/bus/soundwire/devices/sdw:.../dev-status/device_number
> +
> +Date:		September 2020
> +
> +Contact:	Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
> +		Bard Liao <yung-chuan.liao@linux.intel.com>
> +		Vinod Koul <vkoul@kernel.org>
> +
> +Description:	SoundWire Slave status
> +
> +		These properties report the Slave status, e.g. if it
> +		is UNATTACHED or not, and in the latter case show the
> +		device_number. This status information is useful to
> +		detect devices exposed by platform firmware but not
> +		physically present on the bus, and conversely devices
> +		not exposed in platform firmware but enumerated.
> +
>  What:		/sys/bus/soundwire/devices/sdw:.../dev-properties/mipi_revision
>  		/sys/bus/soundwire/devices/sdw:.../dev-properties/wake_capable
>  		/sys/bus/soundwire/devices/sdw:.../dev-properties/test_mode_capable
> diff --git a/drivers/soundwire/slave.c b/drivers/soundwire/slave.c
> index 19b012310c29..ee554a215e44 100644
> --- a/drivers/soundwire/slave.c
> +++ b/drivers/soundwire/slave.c
> @@ -6,6 +6,7 @@
>  #include <linux/soundwire/sdw.h>
>  #include <linux/soundwire/sdw_type.h>
>  #include "bus.h"
> +#include "sysfs_local.h"
>  
>  static void sdw_slave_release(struct device *dev)
>  {
> @@ -83,6 +84,7 @@ int sdw_slave_add(struct sdw_bus *bus,
>  		return ret;
>  	}
>  	sdw_slave_debugfs_init(slave);
> +	sdw_slave_status_sysfs_init(slave);
>  
>  	return ret;
>  }
> diff --git a/drivers/soundwire/sysfs_local.h b/drivers/soundwire/sysfs_local.h
> index ff60adee3c41..52f2eabf5b4b 100644
> --- a/drivers/soundwire/sysfs_local.h
> +++ b/drivers/soundwire/sysfs_local.h
> @@ -8,6 +8,10 @@
>   * SDW sysfs APIs -
>   */
>  
> +/* basic routine to report status of Slave (attachment, dev_num) */
> +int sdw_slave_status_sysfs_init(struct sdw_slave *slave);
> +
> +/* additional device-managed properties reported after driver probe */
>  int sdw_slave_sysfs_init(struct sdw_slave *slave);
>  int sdw_slave_sysfs_dpn_init(struct sdw_slave *slave);
>  
> diff --git a/drivers/soundwire/sysfs_slave.c b/drivers/soundwire/sysfs_slave.c
> index f510071b0add..36eaca343601 100644
> --- a/drivers/soundwire/sysfs_slave.c
> +++ b/drivers/soundwire/sysfs_slave.c
> @@ -16,10 +16,14 @@
>  
>  /*
>   * The sysfs for Slave reflects the MIPI description as given
> - * in the MIPI DisCo spec
> + * in the MIPI DisCo spec. dev-status properties come directly
> + * from the MIPI SoundWire specification.
>   *
>   * Base file is device
>   *	|---- modalias
> + *	|---- dev-status
> + *		|---- status
> + *		|---- device_number
>   *	|---- dev-properties
>   *		|---- mipi_revision
>   *		|---- wake_capable
> @@ -212,3 +216,61 @@ int sdw_slave_sysfs_init(struct sdw_slave *slave)
>  
>  	return 0;
>  }
> +
> +/*
> + * the status is shown in capital letters for UNATTACHED and RESERVED
> + * on purpose, to alert users to the fact that these status values are
> + * not expected.
> + */
> +static const char *const slave_status[] = {
> +	[SDW_SLAVE_UNATTACHED] =  "UNATTACHED",
> +	[SDW_SLAVE_ATTACHED] = "Attached",
> +	[SDW_SLAVE_ALERT] = "Alert",
> +	[SDW_SLAVE_RESERVED] = "RESERVED",
> +};
> +
> +static ssize_t status_show(struct device *dev,
> +			   struct device_attribute *attr, char *buf)
> +{
> +	struct sdw_slave *slave = dev_to_sdw_dev(dev);
> +
> +	return sprintf(buf, "%s\n", slave_status[slave->status]);
> +}
> +static DEVICE_ATTR_RO(status);
> +
> +static ssize_t device_number_show(struct device *dev,
> +				  struct device_attribute *attr, char *buf)
> +{
> +	struct sdw_slave *slave = dev_to_sdw_dev(dev);
> +
> +	if (slave->status == SDW_SLAVE_UNATTACHED)
> +		return sprintf(buf, "%s", "N/A");
> +	else
> +		return sprintf(buf, "%d", slave->dev_num);
> +}
> +static DEVICE_ATTR_RO(device_number);
> +
> +static struct attribute *slave_status_attrs[] = {
> +	&dev_attr_status.attr,
> +	&dev_attr_device_number.attr,
> +	NULL,
> +};
> +
> +/*
> + * we don't use ATTRIBUTES_GROUP here since we want to add a subdirectory
> + * for device-status
> + */
> +static struct attribute_group sdw_slave_status_attr_group = {
> +	.attrs	= slave_status_attrs,
> +	.name = "dev-status",
> +};
> +
> +/*
> + * We can't use devm_ here, otherwise resources would be added before
> + * the driver probe. The group is removed in device_del() however.
> + */
> +
> +int sdw_slave_status_sysfs_init(struct sdw_slave *slave)
> +{
> +	return device_add_group(&slave->dev, &sdw_slave_status_attr_group);

DOesn't this race with userspace?  Why not make this part of the default
set of device attributes and have the driver core create them
automatically?

thanks,

greg k-h