On Thu, Nov 08, 2007 at 11:37:51AM -0800, Chris Poon wrote: > Hello, > > It seems that the kernel package for sparc64 have the loop driver built-in > instead of being a module, which makes loop-aes unusable - can't load either > the loop.ko or the loop-aes.ko, kernel that the module export duplicate > symbol loop_register_transfer (owned by kernel), and trying to mount a > loop-aes swap results in "swapon: LOOP_SET_STATUS failed" as the built-in > loop driver doesn't know what to do with the loop-aes options. > > I assume the kernel ABI will change if the loop driver is compiled as a > module instead but to allow loop-aes to work, it's either a module or > not built it at all (which doesn't make sense if you don't use loop-aes > but still uses the loop driver). Again, I ended up with a custom kernel > without the loop driver and that works, so I'm wondering if the stock > kernel package will be rebuilt to support it?
Seems like a reasonable request; only issue I see is if current etch issues are relying on a built-in loop module for some reason. And yeah, its probably an ABI change, so if we agree to do it, it would probably not happen until 4.0r3. Please file a bug (against linux-2.6) and we can discuss it there. -- dann frazier -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
