On 6 May 2015 at 14:21, Peter Crosthwaite <peter.crosthwa...@xilinx.com> wrote: > On Wed, May 6, 2015 at 6:18 AM, Peter Maydell <peter.mayd...@linaro.org> > wrote: >> On 6 May 2015 at 14:02, Peter Crosthwaite <peter.crosthwa...@xilinx.com> >> wrote: >>> Actually NVM. I got it working on a diff machine. Did you have any >>> thoughts on the GICC mirror issue while I respin? >> >> You mean the thing where the GICC might not be at the bottom of >> a 64K page? Just map it wherever it lives in the hardware you're >> modelling... If your h/w is one of the few that's taken the >> "map the pages mirrored over the page" then a suitable container >> and aliases should handle that I think. >> > > Ok, should the alias be machine model level or an option for gic itself?
It's not really part of the GIC, it's a machine model wiring thing. (This is only relevant if you have an external GIC that's not part of the CPU proper anyway.) -- PMM