[PATCH linux dev-4.7 v2] drivers/fsi: Remove hub devices during master unregister

[Christopher Bostic]

On 3/9/17 4:54 AM, Joel Stanley wrote:
> On Wed, Mar 8, 2017 at 3:16 AM, Christopher Bostic
> <cbostic at linux.vnet.ibm.com> wrote:
>> All hub devices must be removed when the primary FSI master
>> is being unregistered.  This ensures that the system is in
>> a clean state when another scan is attempted.  Failure to do so
>> will result in duplicate /sys device name errors.
>>
>> Signed-off-by: Christopher Bostic <cbostic at linux.vnet.ibm.com>
>> ---
>> v2: Fix missing return in non void function warning
>> ---
>>   drivers/fsi/fsi-core.c | 18 ++++++++++++++++++
>>   1 file changed, 18 insertions(+)
>>
>> diff --git a/drivers/fsi/fsi-core.c b/drivers/fsi/fsi-core.c
>> index bd57b41..bbdf6b8 100644
>> --- a/drivers/fsi/fsi-core.c
>> +++ b/drivers/fsi/fsi-core.c
>> @@ -84,6 +84,7 @@ struct fsi_master_hub {
>>   #define to_fsi_master_hub(d) container_of(d, struct fsi_master_hub, master)
>>   #define to_fsi_slave(d) container_of(d, struct fsi_slave, dev)
>>
>> +static void fsi_master_unscan(struct fsi_master *master);
>>   static int fsi_slave_read(struct fsi_slave *slave, uint32_t addr,
>>                  void *val, size_t size);
>>   static int fsi_slave_write(struct fsi_slave *slave, uint32_t addr,
>> @@ -441,6 +442,7 @@ static int fsi_slave_scan(struct fsi_slave *slave)
>>                          hub->dev.release = hub_master_release;
>>                          hub->master.dev = &hub->dev;
>>                          hub->master.dev->parent = &slave->dev;
>> +                       dev_set_drvdata(&hub->dev, hub);
>>                          rc = device_add(&hub->dev);
>>                          if (rc)
>>                                  return rc;
>> @@ -739,6 +741,20 @@ static int fsi_master_scan(struct fsi_master *master)
>>          return 0;
>>   }
>>
>> +static int fsi_unregister_hubs(struct device *dev, void *data)
>> +{
>> +       struct fsi_master_hub *hub =
>> +                               (struct fsi_master_hub *)dev_get_drvdata(dev);
> dev_get_drvdata returns a void *, so there is no need to cast.
>
> I fixed this up when applying.
>
>> +       if (!hub)
>> +               return 0;
> Is this expected? Should it return an error?
Hi Joel,

Yes it would be expected in cases where the child dev being processed is 
not a hub.  Only hub master has its driver data set. In effect !hub 
means this child is not a hub master.

Thanks
-Chris
>> +
>> +       device_del(dev);
>> +       fsi_master_unscan(&hub->master);
>> +
>> +       return 0;
>> +}
>> +
>>   static void fsi_master_unscan(struct fsi_master *master)
>>   {
>>          struct fsi_slave *slave, *slave_tmp;
>> @@ -756,6 +772,8 @@ static void fsi_master_unscan(struct fsi_master *master)
>>                          device_del(&fsi_dev->dev);
>>                          put_device(&fsi_dev->dev);
>>                  }
>> +               /* Remove any hub masters */
>> +               device_for_each_child(&slave->dev, NULL, fsi_unregister_hubs);
>>                  device_unregister(&slave->dev);
>>          }
>>          master->slave_list = false;
>> --
>> 1.8.2.2
>>