On 2/11/25 16:32, andrzej zaborowski wrote: >> Actually, now that I think about it even more, why would ETRACK or >> EBLOCK access the page itself? They seem superficially like they'd be >> metadata-only too. > I haven't seen a crash in either of these (always in EWB), I didn't > want to imply that. But starting that sequence seems wrong knowing we > cannot reclaim the page.
That's kinda another reason not to delve into the details too deeply. I think you wanted to talk about the "writeback process" as a thing and not really about "ETRACK, EBLOCK and EWB" per se. Writing back an SGX page is the problem. The names of the three instructions that implement the writeback or that there _are_ even three of them isn't super relevant.
