Actually, I think it's ok to set status to 'Rejected' for the bcm43xx-fwcutter package.
This bug is now also registered against linux-source-2.6.20, where I believe the fix would have to be implemented.
My suggestion is that the kernel should be modified to include two parallel versions of bcm43xx. One version would use the softmac stack, the other would use the new devicescape stack.
The kernel could attempt to cut and use V4 firmware and devicescape stack, and if that fails, cut and use version 3 firmware and softmac stack instead.
Actually, I think it's ok to set status to 'Rejected' for the bcm43xx-fwcutter package.
This bug is now also registered against linux-source- 2.6.20, where I believe the fix would have to be implemented.
My suggestion is that the kernel should be modified to include two parallel versions of bcm43xx. One version would use the softmac stack, the other would use the new devicescape stack.
There's even a howto for this on the bcm43xx home page: https:/ /lists. berlios. de/pipermail/ bcm43xx- dev/2006- April/001662. html
The kernel could attempt to cut and use V4 firmware and devicescape stack, and if that fails, cut and use version 3 firmware and softmac stack instead.