On Sun, 26 Oct 2008, Chad Reese wrote: > The Octeon IO space regions are significantly larger than a 32bit kernel > could tlb map easily. The entire range takes 49 bits to address. As a Do you need them all at once? If not, you may be able to get away with ioremap() called on demand. > not particularly clean, but working alternative, we enable 64bit > addressing in the kernel and used XKPHYS to access IO. Every access was > surrounded by a local_irq_save/local_irq_restore. Since this is ugly to > the extreme, maybe we should drop being able to boot a 32bit kernel on > Octeon until something better is worked out. Good idea anyway I would say. With 64-bit processors support for a 32-bit kernel may be a nice feature, but really only if it works straight away with no strange hacks. Otherwise a 64-bit kernel is the obvious choice providing a superset of the functionality, including but not limited to support for the same plain 32-bit (o32) userland if need be. And with the GCC/GAS changes to support -msym32 that happened a while ago the overhead from using a 64-bit kernel has become quite negligible. Maciej
