Re: [PATCH 02/10] driver core: Pull required checks into driver_probe_device()

[Jason Gunthorpe <jgg@nvidia.com>]

On Tue, Jun 08, 2021 at 06:59:10AM +0100, Christoph Hellwig wrote:
> This looks good as-is, but a suggestions for incremental improvements
> below:
> 
> > @@ -1032,10 +1035,10 @@ int device_driver_attach(struct device_driver *drv, struct device *dev)
> >  	__device_driver_lock(dev, dev->parent);
> >  
> >  	/*
> > -	 * If device has been removed or someone has already successfully
> > -	 * bound a driver before us just skip the driver probe call.
> > +	 * If device someone has already successfully bound a driver before us
> > +	 * just skip the driver probe call.
> >  	 */
> > -	if (!dev->p->dead && !dev->driver)
> > +	if (!dev->driver)
> >  		ret = driver_probe_device(drv, dev);
> >  
> >  	__device_driver_unlock(dev, dev->parent);
> 
> It is kinda silly to keep the ->driver check here given that one caller
> completely ignores the return value, and the other turns a 0 return into
> -ENODEV anyway. 

Later patches revise this though, I think you noticed it

> So I think this check can go away, turning device_driver_attach into
> a trivial wrapper for driver_probe_device that adds locking, which
> might be useful to reflect in the naming.

I was scared to change this because "0 on already bound" is uAPI in
sysfs at this point.

And what you noticed in bind_store:

       dev = bus_find_device_by_name(bus, NULL, buf);
       if (dev && dev->driver == NULL && driver_match_device(drv, dev)) {
               err = device_driver_attach(drv, dev);

Seems troublesome as the driver_lock isn't held for that dev->driver
read.

So I'm inclined to delete the above, keep the check in
device_driver_attach and not change the uAPI?

Or the APIs need to be more complicated to support locked and unlocked
callers/etc

Jason