>>>>> "JB" == John Bonesio <[EMAIL PROTECTED]> writes:
Hi, JB> << The Xilinx approach of overwriting the source tree just feels JB> wrong, and no one seems to want to do it that way.>> JB> I am in the group that has control over how this is done. What JB> would you propose be done different? Keep in mind that we are JB> trying to support a process where someone builds a hardware design JB> and the later changes it with new peripherals or perhaps makes JB> minor tweaks. We want to make the updating of the Linux kernel to JB> reflect these hardware changes easy for people. Don't patch the kernel sources from EDK, but instead get the drivers integrated in the mainline kernel. Either do config options for the various IP versions or use runtime detection. We're doing Linux development for several platforms and using a single kernel tree for it all is absolute critical to my sanity.. The EDK drivers are nice as examples of how to use the IP cores, but the defined and documented interface should be the IP, not the EDK drivers. -- Bye, Peter Korsgaard _______________________________________________ Linuxppc-embedded mailing list Linuxppc-embedded@ozlabs.org https://ozlabs.org/mailman/listinfo/linuxppc-embedded
