[Linux-parport] [PATCH] fix parport_serial_pci_resume() ignoring return value from pci_enable_device()

Russell King rmk+lkml at arm.linux.org.uk
Thu Oct 12 19:58:21 EDT 2006

On Fri, Oct 13, 2006 at 01:44:24AM +0200, Jiri Kosina wrote:
> (I guess that the parport_serial_pci_remove() is the right way(tm) to 
> remove the device from the system in non-destructive way even in case 
> pci_enable_device() failed. Tim?)

I suspect all these kind of patches are introducing additional problems.
This one certainly is.  Who's auditing all these patches?  I mean _properly_
auditing them rather than just saying "that's a good idea"?

In this case, you're calling parport_serial_pci_remove() in the failure
path.  That's fine, but this opens the possibility of it being called
twice - once on resume failure and once when the device/driver is
removed.  If this happens, we dereference a NULL pointer. *BAD*.

So, the original without this resume fix is probably far better than
with the fix.

So, patch violently rejected.

Russell King
 Linux kernel    2.6 ARM Linux   - http://www.arm.linux.org.uk/
 maintainer of:  2.6 Serial core

More information about the Linux-parport mailing list