On 08/23/2011 03:53 AM, Ali Bahar wrote:
On Tue, Aug 23, 2011 at 11:35:55AM +0300, Dan Carpenter wrote:
[PATCH 4/5] removes the _netdev_priv() function, but this patch is
still removing call sites. In other words, patch 4/5 breaks the
build.
You're right. Patch 4 won't build by itself. I paid a lot of attention
to this at the outset, but this got missed. Thanks.
So 4 needs to be squashed into 5. Or are you suggesting an
alternative?
To make the point perfectly clear, the code must compile after each patch is
added as a bisection might land in the middle. If possible, it should also work.
I think that all of the _netdev_priv => netdev_priv changes should be in a
separate patch. In addition, reviewing would be a lot easier if the copyright
header changes were not mixed with changes in the code.
Larry Finger's 2010 commit (of the Realtek tarball) seemed to have
been done en masse. So that's what I was going to do, originally. It
looks like this one, too, should be.
The initial inclusion of a staging driver sometimes operates under different
rules. As I recall, I originally was submitting each routine and its header file
in separate patches, but I was asked to combine them to a single patch in the
end. Now that the driver has been around for some time and it is being used,
there is a new rule.
Larry
_______________________________________________
devel mailing list
devel@xxxxxxxxxxxxxxxxxxxxxx
http://driverdev.linuxdriverproject.org/mailman/listinfo/devel
