Re: [git patches] two warning fixes

[Benjamin Herrenschmidt <benh@kernel.crashing.org>]

> Not necessarily as simple as that -- you need to make sure you don't 
> pass something bogus to a sysfs_remove_blah() function at 
> unregister/unload time, if sysfs_create_blah() failed.
> 
> Certainly sysfs_foo() failure is often ignorable in the sense that you 
> want the driver to keep loading... but that does not imply that it is 
> strictly ignorable, if you also consider the associated cleanup code.

It should be trivial enough to have sysfs_create_blah() do enough
initializations before it can fail so that sysfs_remove_blah() do the
right thing regardless.

It's actually a major PITA for a driver that creates a whole bunch of
sysfs files to have to track precisely which ones were created
successfully for the error path. If it's a single function, goto does
the trick but if for some reason it's not, it's really annoying.

Ben.



-------------------------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc.
Still grepping through log files to find problems?  Stop.
Now Search log events and configuration files using AJAX and a browser.
Download your FREE copy of Splunk now >>  http://get.splunk.com/