Jens Nerche wrote: > > >The only time we care is before transferring execution to > >the guest, which is a good time to check and possibly dump > >the vcode cache and rescan. > This doesn't work with self modifying code (yes, that's a bit > braindead to write self mod code, but possible...). Like I said before, ignore the arguments about the _current_ code page. That page is protected by the I&D TLB trick. Since we control the flow of execution to _other_ pages, then it does work for modified code, since any time before allowing execution, we would look at the dirty bit before resuming execution. Since we always keep a copy of the actual data page as it should be, we can always start from scratch. -Kevin
- Re: Another plex86 technical thing Kevin P. Lawton
- RE: Another plex86 technical thing Drew Northup
- Re: Another plex86 technical thing Kevin P. Lawton
- Re: Another plex86 technical thing Jens Nerche
- Re: Another plex86 technical thing Ulrich Weigand
- Re: Another plex86 technical thing Kevin P. Lawton
- Re: Another plex86 technical thing Ramon van Handel
- Re: Another plex86 technical thing Jens Nerche
- Re: Another plex86 technical thing Kevin P. Lawton
- Re: Another plex86 technical thing Jens Nerche
- RE: Another plex86 technical thing Kevin P. Lawton
- RE: Another plex86 technical thing Drew Northup
- Re: Another plex86 technical thing Kevin P. Lawton
- RE: Another plex86 technical thing Ramon van Handel
- Re: Another plex86 technical thing Nick Behnken
- Re: Another plex86 technical thing Ramon van Handel
- Re: Another plex86 technical thing Kevin P. Lawton
- Re: Another plex86 technical thing Jens Nerche
- Re: Another plex86 technical thing Kevin P. Lawton
