So here is what I did in the meantime: * Created a version for 2.4 that puts back all old compatibility stuff that was removed either during the propagation set or the cleanup set. * Created a versio
Same here, but I'd like to have a list somewhere of what each of the ABI versions is for and how they're supposed to behave. It's starting to look like we're going to be adding these on a semi-regul
Where should such a list go ? Currently, 0 or none is for doing everything the old way. 1 is for not setting slaves HW addr via ifenslave and leaving them in down state so the driver gets them with t
Jeff, I'm going to need a ruling from you: We understood from David that support of old ioctl definitions (i.e. those mapped to SIOCDEVPRIVATE) needs to be removed in the 2.6 kernel. This will break
Oh, I was thinking either in a comment block (by the ABI version definitions, for example), tacked on to the end of bonding.txt, or maybe in a README.ABI in the bonding/ directory. I'm not too conce
I say: Toast the whatever_OLD guys (the backwards compat ioctls) in both 2.4 and 2.6. Backwards compatibility is good, but there are limits, and I think these have reached their limit. Keep the ABI
My recommendations are more towards the middle than either end. I would like to see us get rid of the _OLD ioctls in the 2.6 kernel specifically because it uses the SIOCDEVPRIVATE ioctls. This only
So here is what I did in the meantime: * Created a version for 2.4 that puts back all old compatibility stuff that was removed either during the propagation set or the cleanup set. * Created a versio
Same here, but I'd like to have a list somewhere of what each of the ABI versions is for and how they're supposed to behave. It's starting to look like we're going to be adding these on a semi-regul
Where should such a list go ? Currently, 0 or none is for doing everything the old way. 1 is for not setting slaves HW addr via ifenslave and leaving them in down state so the driver gets them with t
Jeff, I'm going to need a ruling from you: We understood from David that support of old ioctl definitions (i.e. those mapped to SIOCDEVPRIVATE) needs to be removed in the 2.6 kernel. This will break
Oh, I was thinking either in a comment block (by the ABI version definitions, for example), tacked on to the end of bonding.txt, or maybe in a README.ABI in the bonding/ directory. I'm not too conce
I say: Toast the whatever_OLD guys (the backwards compat ioctls) in both 2.4 and 2.6. Backwards compatibility is good, but there are limits, and I think these have reached their limit. Keep the ABI
My recommendations are more towards the middle than either end. I would like to see us get rid of the _OLD ioctls in the 2.6 kernel specifically because it uses the SIOCDEVPRIVATE ioctls. This only