On 06/07/16 16:22, Stanislav Shmarov wrote: > > Yes, exactly. > > There is no point for returning to main loop immediately when current > TB is found on host page and is retranslated. We can continue > invalidation of TBs, and finally remove host page write protection. So > there will be no second SEGFAULT. > > And when generating TB for next instructions, host page will be locked > again, if TB includes instructions from that page. >
I think that could work. Regards, Sergey
