Gabriel, Agreed, but the question on the table is ownership of the current code.
- kumar On Jun 8, 2005, at 4:24 AM, Gabriel Paubert wrote: > On Tue, Jun 07, 2005 at 04:47:35PM -0500, Kumar Gala wrote: >> The following boards have no owners and if they dont get one the plan >> is to kill them: >> >> ep405 >> mvme5100 >> pcore > > Actually mvme5100 and pplus could easily be merged with generic > PreP IIRC. At least I wrote a bootloader that remapped my > MVME2400/2600/2700 to look like an MVME5100 (CHRP memory map > but residual data and no OF). This allows much larger > windows to map to the VME bus (which is often very > sparsely populated). > > The only real difference at boot between MVME5100 and > pplus is the mapping of physical addresses to PCI > I/O and memory space, and perhaps also the way to access > the console device. But I don't have any MVME5100 nor > do I plan do buy one (perhaps an MVME6100 but the host > bridge is completely different). > > The question is whether we want a residual data to > pseudo OF device-tree translator in the bootloader > or another way of passing board data to the kernel. > > Gabriel
