On Tue, Nov 22, 2011 at 12:26:21AM -0500, Sean MacLennan wrote: > I have tripped over a bug in my new setup. If you don't build > rtllib.ko as a module, the r8192e driver cannot be a module either. The > opposite is not true because if rtllib is a module, r8192e will always > be a module. > > Currently the r8192e forces a module, that is how I noticed the > problem. There are a few of ways we can go: > > 1) Force rtllib to always be a module. This is the cleanest solution > since you cannot produce a non-compiling driver. > > 2) Remove the requirement that r8192e always be a module and rely on the > user to get the settings right. > That won't work. > 3) Magic Kconfig fu that gets r8192e driver right. > The magic you are looking for is: "depends on m"? The original driver used this so it should be fine for the broken out ones as well. > 4) Modify the driver to work right in the mixed case. I don't want to > do this only because it fixes an edge case only hit by people who > use all config yes or config random. Those are very important people and we don't want to annoy them. Can you redo the patchset, and move that one export from patch 2 to patch 1? regards, dan carpenter
Attachment:
signature.asc
Description: Digital signature
_______________________________________________ devel mailing list devel@xxxxxxxxxxxxxxxxxxxxxx http://driverdev.linuxdriverproject.org/mailman/listinfo/devel
