Re: [PATCH] cxl: Fix unbalanced pci_dev_get in cxl_probe

On Wed, 2015-09-09 at 16:56 +1000, Daniel Axtens wrote:

Ahaha so I was wrong, device_add does grab a reference. 

quoted

Currently, cxl_probe(pdev):
 1) calls pci_dev_get(pdev)
 2) calls cxl_adapter_init
    a) init calls cxl_adapter_alloc, which creates a struct cxl, 
       conventionally called adapter. This struct contains a 
       device entry, adapter->dev.

    b) init calls cxl_configure_adapter, where we set 
       adapter->dev.parent = &dev->dev (here dev is the pci dev)

So at this point, the cxl adapter's device's parent is the pci device
that I want to be refcounted.

    c) init calls cxl_register_adapter (which inexplicably is in file.c)

       *) cxl_register_adapter calls device_register(&adapter->dev) 

So now we're in device_register, where dev is the adapter device, and we
want to know if the PCI device is safe after we return.

device_register(&adapter->dev) calls device_initialize() and then
device_add().

device_add() does a get_device(). That ends up being a kobject_get() on
the adapter device kobj, which will increment the kref on the adapter
device. 

I was right up to this point, but I didn't read enough of device_add.

device_add explicitly grabs the device's parent, and calls get_device on
it:

         parent = get_device(dev->parent);

So it turns out we *are* protected against the device disappearing, my
patch is correct and I don't need a v2.

Thanks to Ian for making me recheck device_add :)

Thanks for digging into it.

Do you mind massaging that explanation into something I can put into the
changelog?

cheers