Re: klists and struct device semaphores

[Alan Stern]

On Wed, 6 Apr 2005, Patrick Mochel wrote:

> > Third, why does device_release_driver() call klist_del() instead of
> > klist_remove() for dev->knode_driver?  Is that just a simple mistake?
> > The klist_node doesn't seem to get unlinked anywhere.
> 
> It can be called from driver_for_each_device() when the driver has been
> unloaded. Since that increments the reference count for the node when it's
> unregistering it, klist_remove() will deadlock. Instead klist_del() is
> called, and when the next node is grabbed, that one will be let go and
> removed from the list.

The patch looks good.  But isn't there still a problem with
device_release_driver()?  It doesn't wait for the klist_node to be removed
from the klist before unlocking the device and moving on.  As a result, if
another driver was waiting to bind to the device you would corrupt the
list pointers, by calling klist_add_tail() for the new driver before
klist_release() had run for the old driver.

I'll be interested to see how you manage to solve this.  The only way I 
can think of is to avoid using driver_for_each_device() in 
driver_detach().

Alan Stern

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/