> Ah, yes... I used grep to see that there was a rduniq entry in decoder.isa > but failed to realize that that was the SE-mode-only version. The funny thing is that even though I rightly pointed that out, I still was confusing the callpal rduniq and the MTPR/MFPR. Seems that we could have MTPR be serializing and MFPR not be, though we may in fact need some mechanism to determine which if any of the MFPRs have side effects. From a quick glance of readIPR, it seems that none do. That said, there's a panic in there on some write only registers and we probably need to turn that into an opdec fault somehow so misspeculative accesses don't crash the simulator.
> It does seem that having the serialization decision made based on the > particular IPR is necessary. I see that call_pal is non-speculative, which > could cause problems even if the IPR accesses are fixed. I wonder why the callpal itself needs to be non-speculative. Does it have to do with switching modes? The mode is in the PC, so it would seem to me to be renamed already. Nate _______________________________________________ m5-users mailing list [email protected] http://m5sim.org/cgi-bin/mailman/listinfo/m5-users
