On Fri, Jun 19, 2009 at 6:18 PM, Carl-Daniel Hailfinger<[email protected]> wrote:
> > True, because it does not use locking. v3 should be easy to convert to > locking which actually works. we're back where we started. The question is, is there a locking construct that will work on CAR in common for every CPU. I've come to believe that's a lot harder than I thought it was. You also pointed this problem out and were looking at using other means -- cpu registers, etc -- to achieve the goal of locking. The locking for the RAM part is a solved problem, and it works in v2, so I see no need to revisit that question. The question is the locking for the ROM code, given all the different types of CAR we deal with. Do you have a solution to that problem? Let's start there first. ron -- coreboot mailing list: [email protected] http://www.coreboot.org/mailman/listinfo/coreboot
