On 04.06.2011, at 01:42, Scott Wood wrote: > On Thu, 2 Jun 2011 20:37:30 +0200 > Alexander Graf <ag...@suse.de> wrote: > >> + case MPC8544_GUTS_ADDR_PVR: >> + value = env->spr[SPR_PVR]; >> + break; >> + case MPC8544_GUTS_ADDR_SVR: >> + value = env->spr[SPR_E500_SVR]; >> + break; > > Heh, I didn't even realize these were in there -- I was just thinking of > the guest using the SPR version of SVR to decide how to program devices > like guts or tsec, and the conflicting demands that might place on that > value if you have both directly-assigned devices and emulated devices that > target a specific chip. > > I can see where that produced confusion. :-)
Yeah, pretty confusing :). Well, they were easy enough to add and at least cover the very unusual use-case that a guest OS reads from mmio instead of its SPRs *shrug* :). > >> + default: >> + fprintf(stderr, "Unknown register read: %x = %x\n", (int)addr, >> value); >> + break; >> + } > > Should say something like "guts" or __func__ in there somewhere. > Especially if we're going to throw away the upper 4 bits of the address. :-) Good point :). v3 is out. Alex
